Microfluidics for Biological Applications
Wei-Cheng Tian · Erin Finehout
Editors
Microfluidics for Biological Applications
13
Editors Wei-Cheng Tian General Electric Global Research Center 1 Research Circle Niskayuna, NY 12309
Erin Finehout General Electric Global Research 1 Research Circle Niskayuna, NY 12309
ISBN: 978-0-387-09479-3
e-ISBN: 978-0-387-09480-9
Library of Congress Control Number: 2008930844 2008 Springer Science+Business Media, LLC All rights reserved. This work may not be translated or copied in whole or in part without the written permission of the publisher (Springer Science+Business Media, LLC, 233 Spring Street, New York, NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed is forbidden. The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights. Printed on acid-free paper springer.com
To my family Erin Finehout
To Gan-Wu, Yu-Hsien, Wei-Hua, Kaitlyn, Darren, and Jennifer Wei-Cheng Tian
Preface
In Nobel Prize winner Richard Feynman’s well-known 1959 speech “There’s Plenty of Room at the Bottom” [1] he marvels that although many biological systems, such as cells, are very small, they are active and perform a number of functions. He then poses the challenge “Consider the possibility that we too can make a thing very small which does what we want – that we can manufacture an object that maneuvers at that level!” [1] In this book we hope to show readers that we are getting closer to meeting this challenge. We have tools to manipulate and analyze small volumes of biomolecules (such as DNA and protein); we can manipulate and analyze individual cells; and we can create nanodrops, the size of a cell, to perform specific chemical reactions. All of these have been made possible by the application of microfluidics. This book consists of a selection of review articles that are intended to show how microfluidics is applied to solve biological problems; why microfluidics continues to play an important role in this field; and what needs to be done next. We will introduce not only the various technologies of microfluidics but also how to link these technologies to different biological applications at the industrial and academic level. Chapters 1-3 give perspective on the history and development of microfluidic technologies. They also serve to give a physical understanding of microfluidic devices. Chapter 1 covers the physics and fluid dynamics of microscale flows. Chapter 2 summarizes the materials and methods used to fabricate microfluidic devices in biological applications. Chapter 3 gives solutions to how these microscale devices can be interfaced with the macro scale world. Chapters 4-6 give overviews on how microfluidic systems have been used to study and manipulate specific classes of components. Microfluidic devices have been used to: prepare, amplify, and analyze DNA samples (Chapter 4); separate and analyze protein mixtures (Chapter 5); and culture, separate, and analyze cells (Chapter 6). Chapters 7-10 focus on specific biological applications of microfluidics: tissue engineering (Chapter 7), high throughput screening (Chapter 8), diagnostics (Chapter 9), and biodefense (Chapter 10). And finally, Chapter 11 discusses
viii
emerging trends in the microfluidics field and the current challenges to the growth and continuing success of the field. In all the chapters, the authors give information on the biological problems that need to be solved, the current research that is being done to address them, and the obstacles that still remain. In addition, there are summaries of the types of products that have been commercialized in each area. This book is intended to be used at the senior undergraduate or graduate level for students. It will also be a great resource for researchers and scientists in the biotechnology, pharmaceutical, and life science industries. We hope to provide the readers with an overview of microfluidics and its current applications to encourage readers to think about how these technologies could help them in their own fields. Reading through the chapters there are a few recurring themes that merit being mentioned here. The first is that the application of microfluidics isn’t just about saving time, cutting costs, and needing less reagents. Working in the microfluidic regime enables scientists to perform experiments and use techniques that simply aren’t possible at a larger scale. The second theme is that for the microfluidics field as a whole to continue to move forward in the biological area, it is vital that scientists from different fields (engineers, chemists, material scientists, biologist, etc.) work together. Only with such collaborations can one be sure that the right questions are being addressed, the right methods are being applied, and the optimal tools are being used. The editors would like to thank Steven Elliot and Angela DePina at Springer for their help in pulling this book together. We would also like to show our appreciation to the authors for all of the time and effort they put towards writing their chapters. Lastly, we’d like to thank our friends and family for their support and patience during this project.
References: 1. Feynman RP (1960) There’s plenty of room at the bottom: An invitation to enter a new field of physics. Engineering and Science 23:22-36.
Contents
Chapter 1 Introduction to Microfluidics ................................................ 1 Abstract................................................................................................... 1 1 Introduction to Microfluidics............................................................... 2 1.2 History of Microfluidics ................................................................... 3 1.2.1 The beginning: Gas chromatography and capillary electrophoresis.................................................................................... 3 1.2.2 The microfluidic advantage ....................................................... 5 1.2.3 Modular separation, reaction and hybridization systems .......... 7 1.2.4 Integrated systems ..................................................................... 8 1.3 Fluidics and Transport Fundamentals............................................. 10 1.3.1 The continuum approximation................................................. 10 1.3.2 Laminar flow ........................................................................... 10 1.3.3 Diffusion in microfluidic systems ........................................... 12 1.3.4 Surface forces and droplets...................................................... 14 1.3.5 Pumps and valves .................................................................... 16 1.3.6 Electrokinetics ......................................................................... 16 1.3.7 Thermal management .............................................................. 18 1.4 Device Fabrication.......................................................................... 18
x
1.4.1 Materials .................................................................................. 19 1.4.2 Fabrication and assembly ........................................................ 20 1.5 Biological Applications .................................................................. 21 1.5.1 Genetic analysis (DNA/RNA) ................................................. 22 1.5.2 Proteomics ............................................................................... 22 1.5.3 Cellular assays ......................................................................... 23 1.5.4 Drug delivery and compatibility.............................................. 24 1.6 The Future....................................................................................... 26 1.6.1 Potential demand/market for microfluidic devices.................. 26 1.6.2 Current products ...................................................................... 27 1.6.3 Challenges and the future ........................................................ 28 References ............................................................................................ 29 Chapter 2 Materials and Microfabrication Processes for Microfluidic Devices ...................................................................................................... 35 Abstract................................................................................................. 35 2.1 Introduction .................................................................................... 36 2.2 Silicon Based Materials .................................................................. 37 2.2.1 Micromachining of silicon....................................................... 39 2.2.2 Bulk micromachining .............................................................. 39 2.2.3 Surface micromachining.......................................................... 46 2.3 Glass Based Materials..................................................................... 49 2.3.1 Microfabrication in glass ......................................................... 51 2.4 Wafer Bonding ............................................................................... 56 2.4.1 Fusion bonding ........................................................................ 57 2.4.2 Anodic bonding ....................................................................... 57 2.4.3 Adhesive bonding.................................................................... 58 2.5 Polymers ......................................................................................... 59 2.5.1 Microfabrication ...................................................................... 59 2.5.2 Polymer materials:................................................................... 64 2.6 Conclusion ...................................................................................... 82 References ............................................................................................ 82 Chapter 3 Interfacing Microfluidic Devices with the Macro World.. 93 Abstract................................................................................................. 93 3.1 Introduction .................................................................................... 94 3.2 Typical Requirements for Microfluidic Interfaces ......................... 94 3.3 Review of Microfluidic Interfaces.................................................. 95 3.3.1 World-to-chip interfaces.......................................................... 95 3.3.2 Chip-to-world interfaces........................................................ 103 3.4. Future Perspectives...................................................................... 112 References .......................................................................................... 113
xi
Chapter 4 Genetic Analysis in Miniaturized Electrophoresis Systems.....................................................................................................117 Abstract............................................................................................... 117 4.1 Introduction .................................................................................. 118 4.1.1 Status of genetic analyses ...................................................... 118 4.1.2 Genetic analysis by miniaturized electrophoresis system ..... 119 4.2 Microchip Electrophoresis for Genomic Analysis........................ 122 4.2.1 Material and fabrication of electrophoresis microchips ........ 123 4.2.2 Theory of gel electrophoresis of DNA .................................. 125 4.2.3 Gel matrices........................................................................... 126 4.2.4 Novel DNA separation strategies on microchips................... 130 4.2.5 Surface coating methods for microchannel walls.................. 134 4.3 Parallelization in Microchip Electrophoresis................................ 137 4.4 Integration in Microchip Electrophoresis for Genetic Analysis ... 139 4.4.1 Sample preparation on microchip.......................................... 139 4.4.2 System integration ................................................................. 141 4.5 Commercial Microfluidic Instruments for Genetic Analyses....... 144 4.5.1 Commercial microchip electrophoresis instruments for genetic analysis ........................................................................................... 145 4.5.2 Integrated microfluidic instruments for genetic analyses ...... 147 4.6 Microfluidic Markets and Future Perspectives............................. 150 References .......................................................................................... 151 Chapter 5 Microfluidic Systems for Protein Separations ................. 165 Abstract............................................................................................... 165 5.1 Introduction .................................................................................. 166 5.1.1 Advantages of microfluidic chips for protein separations..... 166 5.1.2 Limitations of microfluidic chips in proteomics applications167 5.1.3 Substrates used for proteomic analysis.................................. 167 5.2. Microfluidic Chips for Protein Separation................................... 168 5.2.1 Microchip-based electrophoretic techniques ......................... 169 5.2.2 Microchip chromatography ................................................... 172 5.3 Integrated Analysis in Microchips................................................ 175 5.3.1 Integration of sample preparation with analysis.................... 175 5.3.2 Multi-dimensional separation in microchips ......................... 177 5.3.3 Chips integrated with mass spectrometry .............................. 180 5.4. Future Directions ......................................................................... 180 References .......................................................................................... 181 Chapter 6 Microfluidic Systems for Cellular Applications............... 185 Abstract............................................................................................... 185 6.1 Introduction .................................................................................. 186 6.1.1 Physiological advantages....................................................... 188
xii
6.1.2 Biological advantages............................................................ 189 6.1.3 Economical advantages ......................................................... 191 6.2 Microfluidic Technology for Cellular Applications ..................... 191 6.2.1 Microfluidic cell isolation/separation .................................... 191 6.2.2 Microfluidic cell culture ........................................................ 200 6.2.3 Microfluidic cell analysis ...................................................... 208 6.3 Commercialization of Microfluidic Technology .......................... 211 6.4 Concluding Remarks .................................................................... 214 References .......................................................................................... 215 Chapter 7 Microfluidic Systems for Engineering Vascularized Tissue Constructs............................................................................................... 223 Abstract:.............................................................................................. 224 7.1 Introduction .................................................................................. 224 7.2 Generating 2D Vascularized Tissue Constructs Using Microfluidic Systems............................................................................................... 226 7.3 Generating 3D Vascularized Tissue Constructs Using Microfluidic Systems............................................................................................... 230 7.4 Hydrogel-based Microfluidic Systems for Generating Vascularized Tissue Constructs................................................................................ 232 7.5 Mathematical Modeling to Optimize the Microfluidic Systems for Generating Vascularized Tissue Constructs ....................................... 235 7.6 Future Challenges ......................................................................... 237 7.7 Conclusions .................................................................................. 237 References .......................................................................................... 237 Chapter 8 High Throughput Screening Using Microfluidics............ 241 Abstract............................................................................................... 241 8.1 Introduction .................................................................................. 242 8.2 Cell-Based Assays ........................................................................ 244 8.2.1 High throughput cell culture.................................................. 245 8.2.2 Cell sorting for high throughput applications........................ 252 8.3 Biochemical Assays...................................................................... 254 8.3.1 PCR ....................................................................................... 254 8.3.2 Electrophoresis ...................................................................... 255 8.3.3 Others .................................................................................... 255 8.4 Drug Screening Applications........................................................ 258 8.5 Users and Developers of µF HTS Platforms ................................ 259 8.5.1 Users: Research labs, academic screening facilities, and pharmaceutical................................................................................ 260 8.5.2 Commercialized products in HTS ......................................... 261 8.6 Conclusion .................................................................................... 262
xiii
8.7 Acknowledgements ...................................................................... 263 References .......................................................................................... 263 Chapter 9 Microfluidic Diagnostic Systems for the Rapid Detection and Quantification of Pathogens .......................................................... 271 Abstract............................................................................................... 271 9.1 Introduction .................................................................................. 272 9.1.1 Infectious pathogens and their prevalence............................. 272 9.1.2 Traditional pathogen detection methods................................ 274 9.1.3 Microfluidic techniques......................................................... 276 9.2 Review of Research ...................................................................... 277 9.2.1. Pathogen detection/quantification techniques based on detecting whole cells ...................................................................... 277 9.2.2 Pathogen detection/quantification techniques based on detecting metabolites released or consumed................................... 294 9.2.3 Pathogen detection/quantification through microfluidic immunoassays and nucleic acid based detection platforms............ 297 9.3 Future Research Directions........................................................... 305 References .......................................................................................... 307 Chapter 10 Microfluidic Applications in Biodefense......................... 323 Abstract............................................................................................... 323 10.1 Introduction ................................................................................ 324 10.2 Biodefense Monitoring ............................................................... 326 10.2.1 Civilian biodefense .............................................................. 326 10.2.2 Military biodefense.............................................................. 328 10.3 Current Biodefense Detection and Identification Methods ........ 330 10.3.1 Laboratory detection............................................................ 331 10.3.2 Field detection ..................................................................... 332 10.4 Microfluidic Challenges for Advanced Biodefense Detection and Identification Methods........................................................................ 333 10.5 Microscale Sample Preparation Methods ................................... 335 10.5.1 Spore disruption................................................................... 336 10.5.2 Pre-separations .................................................................... 336 10.5.3 Nucleic acid purifications.................................................... 337 10.6 Immunomagnetic Separations and Immunoassays ..................... 339 10.6.1 Immunomagnetic separations .............................................. 340 10.6.2 Immunoassays ..................................................................... 341 10.7 Proteomic Approaches................................................................ 345 10.8 Nucleic Acid Amplification and Detection Methods ................. 346 10.8.1 PCR and qPCR detection of pathogens for biodefense ....... 347 10.8.2 Miniaturized and Microfluidic PCR .................................... 348
xiv
10.8.3 Heating and cooling approaches.......................................... 349 10.8.4 Miniaturized PCR and qPCR for biodefense....................... 350 10.8.5 Other Nucleic acid amplification methods .......................... 351 10.9 Microarrays................................................................................. 352 10.9.1 Microarrays and microfluidics............................................. 353 10.10 Microelectrophoresis and Biodefense....................................... 354 10.10.1 Microelectrophoresis technologies .................................... 356 10.11 Integrated lab-on-a-chip systems and biodefense ..................... 358 10.11.1 Full microfluidic integration for biodefense...................... 363 10.12 Summary and Perspectives ....................................................... 363 References .......................................................................................... 365 Chapter 11 Current and Future Trends in Microfluidics within Biotechnology Research ........................................................................ 385 Abstract............................................................................................... 385 11.1 The Past – Exciting Prospects..................................................... 386 11.2 The Present – Kaleidoscope-like Trends .................................... 388 11.2.1 Droplet microfluidics........................................................... 389 11.2.2 Integrating Active Components in Microfluidics ................ 391 11.2.3 Third world - paper microfluidics – George Whitesides ..... 394 11.2.4 Microfluidic solutions for enhancing existing biotechnology platforms......................................................................................... 395 11.2.5 Microfluidics for cell biology – seeing inside the cell with molecular probes ............................................................................ 400 11.2.6 Microfluidics for cell biology – high throughput platforms 401 11.3 The Future – Seamless and Ubiquitous MicroTAS .................... 403 References .......................................................................................... 405 Index........................................................................................................ 413
List of Contributors
Beebe, David J., Ph.D University of Wisconsin–Madison Madison, WI 53706 USA Borenstein, Jeffrey, Ph.D. Draper Laboratory Cambridge, MA 02139 USA Burns, Mark A., Ph.D. University of Michigan Ann Arbor, MI 48109 USA Chang, Dustin S. University of Michigan Ann Arbor, MI 48109 USA Chang, Hsueh-Chia, Ph.D. Center for Microfluidics & Medical Diagnostics Notre Dame, IN 46556 USA University of Notre Dame Notre Dame, IN 46556 USA Chung, Yao-Kuang University of Michigan Ann Arbor, MI 48109 USA
xvi
Cropek, Donald, Ph.D. U.S. Army Corps of Engineers Champaign, IL 61822 USA Du, Yanan, Ph.D. Massachusetts Institute of Technology Cambridge, MA 02139 USA Harvard Medical School Cambridge, MA 02139 USA Gordon, Jason E., Ph.D. Midwest Research Institute Kansas City, MO 64110 USA Horn, Joanne, Ph.D. Microchip Biotechnologies Inc. 6693 Sierra Lane, Suite F, Dublin, CA 94568 USA Jain, Akshat University of Michigan Ann Arbor, MI 48109 USA Jovanovich, Stevan, Ph.D. Microchip Biotechnologies Inc. 6693 Sierra Lane, Suite F, Dublin, CA 94568 USA Khademhosseini, Ali, Ph.D. Massachusetts Institute of Technology Cambridge, MA 02139 USA Harvard Medical School Cambridge, MA 02139 USA
xvii
Kuo, Chuan-Hsien University of Michigan Ann Arbor, MI 48109 USA Langelier, Sean M. University of Michigan Ann Arbor, MI 48109 USA Lee, Abraham P., Ph.D. University of California at Irvine Irvine, CA 92697 USA Micro/nano Fluidics Fundamental Focus (MF3) Center Irvine, CA 92697 USA Lin, Gisela, Ph.D. Micro/nano Fluidics Fundamental Focus (MF3) Center Irvine, CA 92697 USA Mofrad, Mohammad R. Kaazempur, Ph.D. University of California at Berkeley Berkeley, CA 94720 USA Noori, Arash McMaster University Hamilton ON L8S 4L7 CANADA Park, Jihyang University of Michigan Ann Arbor, MI 48109 USA Puccinelli, John P. University of Wisconsin–Madison Madison, WI 53706 USA
xviii
Rhee, Minsoung University of Michigan Ann Arbor, MI 48109 USA Selvanganapathy, P. Ravi, Ph.D. McMaster University Hamilton ON L8S 4L7 CANADA Sengupta, Shramik, Ph.D. University of Missouri Columbia, MO 65211 USA Shaikh, Kashan A., Ph.D. GE Global Research 1 Research Circle, Niskayuna NY 12309 USA Sommer, Greg J., Ph.D University of Michigan Ann Arbor, MI 48109 USA Takayama, Shuichi, Ph.D. University of Michigan Ann Arbor, MI 48109 USA Tavana, Hossein, Ph.D. University of Michigan Ann Arbor, MI 48109 USA Upadhyaya, Sarvesh McMaster University Hamilton ON L8S 4L7 CANADA
xix
Wang, Fang University of Michigan Ann Arbor, MI 48109 USA Wang, Hong, Ph.D. Louisiana State University Baton Rouge, LA 70803 USA Wang, Xuefeng, Ph.D. GE Global Research 1 Research Circle, Niskayuna NY 12309 USA Weinberg, Eli J., Ph.D. Draper Laboratory Cambridge, MA 02139 USA Zeitoun, Ramsey I. University of Michigan Ann Arbor, MI 48109 USA Zhu, Li, Ph.D. GE Global Research 1 Research Circle, Niskayuna NY 12309 USA
Chapter 1 Introduction to Microfluidics
Greg J. Sommer, Dustin S. Chang, Akshat Jain, Sean M. Langelier, Jihyang Park, Minsoung Rhee, Fang Wang, Ramsey I. Zeitoun, and Mark A. Burns University of Michigan, Ann Arbor, MI 48109 Correspondence should be addressed to: Mark Burns (
[email protected])
Keywords: microfluidics, history, fundamentals, applications, commercialization
Abstract Microfluidics – the manipulation and analysis of minute volumes of fluid has emerged as a powerful technology with many established and relevant applications within the biological sciences. Over three decades of research has yielded a wealth of techniques for improving biological assays through both the miniaturization of existing methods, as well as the development of novel analytical approaches. In this introductory chapter we provide an overview of microfluidic technology, beginning with a historical look at the field’s origins. We also present brief synopses of the fundamental physical phenomena driving microfluidics, the techniques employed in device fabrication, and biological applications that have benefited from microfluidic implementation. Finally we conclude with an outlook to the fu-
2
Sommer et. al
ture of the field as microfluidic technology shifts from research laboratories into commercial ventures. This chapter is meant to familiarize readers who may be new to the field of microfluidics, while highlighting areas that will be explored in more detail throughout the text.
1 Introduction to Microfluidics Microfluidic technology has evolved over the past few decades from a molecular analysis endeavor aimed at enhancing separation performance through reduced dimensions, into a diverse field influencing an everexpanding range of disciplines. Microfluidic techniques are being employed in chemistry, biology, genomics, proteomics, pharmaceuticals, biodefense, and other areas where its inherent advantages trump standard methodologies. From a biological standpoint, microfluidics seems especially relevant considering that most biological processes involve small-scale fluidic transport at some point. Examples stem from molecular transfer across cellular membranes, to oxygen diffusivity through the lungs, to blood flow through microscale arterial networks. Microfluidics can also provide more realistic in vitro environments for small-scale biological species of interest. Figure 1.1 provides comparative length scales for several biological structures, as well as common micro-fabrication structures used in microfluidic and MEMS technology.
Fig. 1.1 Approximate length scales for several biological and micro-fabrication structures.
Introduction to Microfluidics
3
In this chapter we will provide a brief introduction to microfluidics, beginning with a historical perspective of the field’s origins and concluding with an outlook on the future. We also present some fundamental transport principles, materials and fabrication basics, and specific biological applications with the aim of familiarizing the reader with important microfluidicrelated concepts and highlighting areas that will be explored further throughout the text.
1.2 History of Microfluidics 1.2.1 The beginning: Gas chromatography and capillary electrophoresis The mid-20th century saw explosive growth in the applicability of chromatography: a technique that revolutionized the field of separation sciences by exploiting molecular distributions between mobile and stationary phases within a column. Theoretical work by Golay [1] on gas chromatography (GC) and van Deemter [2] on liquid chromatography established scaling arguments showing that improved performance could be achieved by reducing open column diameters and packed column particle sizes. Thus columns began being fabricated from fused silica capillaries with diameters on the order of micrometers. Around the same time, capillary electrophoresis (CE) was gaining popularity as a method to separate charged biomolecules. Here too, small bore capillaries proved advantageous as the larger surface area-to-volume ratio allowed for higher applied electric fields and, therefore, improved separation performance. But the roots of microfluidics truly lie in the microelectronics industry. As chemists and biologists were searching for means to further miniaturize their analytical methods, the microelectronics industry was improving its silicon-based micromachining processes using photolithography, etching, and bonding techniques [3]. The merging of the bioanalytical and microelectronics disciplines can be considered the birth of microfluidics. The first silicon-based analysis system was published in 1979, in which Terry et al. from Stanford University fabricated a miniature GC air analyzer on a silicon wafer [4] (Fig. 1.2). However, it was the seminal works by Manz and others in the early 1990’s that demonstrated the microfluidic potential for addressing issues facing analytical methods, and spawned the term micro Total Analysis Systems (µTAS) [5-7].
4
Sommer et. al
Fig. 1.2 Miniature gas chromatograph developed by Terry et al on a silicon wafer in 1979. [8] - Reproduced by permission of The Royal Society of Chemistry.
While researchers continued miniaturizing gas and liquid chromatography columns [6, 9], many of the first successful microfluidic separation devices employed electrophoretic techniques due to the relative simplicity of applying an electric potential to a microchannel versus a high-pressure source such as those required for high pressure liquid chromatography (HPLC). In 1992, Manz et al. demonstrated the first on-chip CE system and initiated the new microfluidic era in separation sciences. That same year, Mathies et al.proposed high-throughput electrophoretic sequencing on arrays of microfluidic devices [10]. In 1993, Harrison et al. demonstrated a micro-CE system in glass which could separate amino acids with about 75,000 theoretical plates in 15 seconds [11]. The next year, Woolley and Mathies successfully miniaturized a microfluidic capillary gel electrophoresis system for DNA analysis, which boasted separation times in as little as 120 seconds [12]. The microfluidics boom had begun. The mid 1990s brought many new concepts and devices as researchers began to investigate microfluidic uses for not only separations, but other applications as well. So why all the ruckus? In the next section we will highlight some advantages that microfluidics can provide over conventional, macroscale methodologies.
Introduction to Microfluidics
5
1.2.2 The microfluidic advantage Silicon micromachining enabled fabrication of channels and features with precision on the order of 1µm. This technological feat enabled the manipulation of micro- (10-6) to atto- (10-18) liter volumes of fluid. Such control brings several advantages from both analytical and economic viewpoints. Here we briefly outline those advantages, while noting that many of these concepts will be further explored throughout this chapter and text. A summary of the advantages are listed in Table 1.1. Table 1.1 Summary of advantages attained with microfluidic systems Microfluidic Advantage Description Less sample and reagent Microfluidic devices typically require 102 – 103 less sample volume than conventional assays. consumption Enhanced heat transfer
Higher surface area-to-volume ratio of microfluidic channels increases effective thermal dissipation.
Faster separations
Higher E-fields results in faster sample migration.
Laminar flow
Low Reynolds number flows reduce sample dispersion.
Electrokinetic manipulation
Electroosmotic flow enables fluid pumping with flat "plug-like" velocity profiles solely via applied E-fields.
Lower power consumption
Fewer components and enhanced thermal dissipation require less power input.
Parallelization
Several assays can be “multiplexed”, or run in parallel on a single chip.
Portability
System integration and reduced power allows for assays to be conducted using portable, hand-held device.
Improved separation efficiency
Efficiency in electrophoretic and chromatographic separations (i.e. number of theoretical plates) proportional to L/d.
6
Sommer et. al
Several measures of analytical performance can be improved through miniaturization. Perhaps one of the most obvious advantages of smaller channel sizes is reduced reagent consumption, leading to less waste and more efficient assays. Reduced reagent consumption becomes especially advantageous for many biological applications where reagents can be very expensive (e.g. antibodies), and sample volumes are often limited. Additionally, the separation efficiency (i.e. number of theoretical plates) of chromatographic and electrophoretic systems is proportional to L/d: the length of the separation channel over its diameter. Therefore long and narrow channels enable improved peak-peak resolution. Because they are so narrow, microfluidic channels also boast flows with very low Reynolds numbers: often Re < 1, meaning the flow is laminar. Such laminar flows inhibit additional dispersion from affecting the band width of a separated plug. Diffusion, however, is more prominent at smaller scales and can be advantageous for mixing applications where, despite very laminar flow, mixing can occur solely via diffusion. Narrow channels also dissipate heat more efficiently, allowing for higher electric fields in electrophoretic systems without adverse Joule heating effects on separation efficiency. As a result the assays will require less time as higher electric fields lead to faster separations. Microfluidic devices often achieve fluid transport with few mechanical components, which can significantly reduce an assay’s complexity and power consumption over its macroscale counterpart. For example, electroosmosis is a process in which bulk electrolytic fluid in a channel is dragged via viscosity by migrating ions near an inherently charged channel wall under the application of an electric field. Electroosmosis allows for: a) bulk fluid “pumping” using only electric fields, thereby eliminating any moving parts, and b) a plug-like, non-parabolic fluid velocity profile that eliminates dispersion caused by parabolic pressure-driven flow. Bulk fluid transport has also been demonstrated on microfluidic devices using other pumping techniques such as capillary wicking, evaporation, thermal gradients, and chemically-induced flow. With many conventional assays it is possible to integrate all analytical steps (sample loading, rinsing, reactions, separation, detection, etc.) into a single, fully-automated platform. Such integration reduces necessary human involvement, potential environmental contamination, and analysis time. With µTAS, or lab-on-a-chip (LOC) devices, such integration can greatly reduce the cost per analysis while providing high throughput through parallelized or multiplexed devices. They can also be potentially integrated into a portable, hand-held format for a variety of point-of-care
Introduction to Microfluidics
7
(POC) applications where proper laboratory access is not available or rapid analysis time is required, including bedside patient care, military and border patrol, and global healthcare scenarios. As fabrication procedures become more standardized, the cost per chip will decrease enabling the production of inexpensive, single-use, disposable chips.
1.2.3 Modular separation, reaction and hybridization systems Along with the initial demonstrations of microfluidic separation systems in the early 1990’s, researchers began exploring other methods with which to fill the “analytical toolbox” necessary to build the envisioned integrated systems. Further motivation for the microfluidics community arose from the explosion of genomics in the 1990’s. Biologists were increasingly interested in exploring DNA and decoding genes and chromosomes, as evidenced by the appeal and success of the Human Genome Project. Therefore much of the early research was directed towards DNA amplification, hybridization and sequencing. One technology that naturally found its way into microfluidic devices was that of microarrays. Microarrays, in which minute biological samples are immobilized as individual spots that may hybridize with an introduced sample, allow for extremely large numbers of parameters to be screened at any one time. In 1991, the use of standard lithography to pattern an array was first introduced by Fodor et al. [13] (Fig. 1.3). Since then, microfabricated arrays have found a home using microfluidic technologies. Microarrays have been developed to pack a maximum amount of DNA strands into a minimal amount of space. As early as 1995, a 96 microwell array was used to detect an organism’s gene expressions [14]. A year later, a DNAchip which had 1046 different DNA strands was developed demonstrating the potential power of microfabricating DNA analysis devices [15].
8
Sommer et. al
Fig. 1.3 DNA microarray fabricated lithographically by Fodor et al in 1991. Each square represents one gene sequence and is 50µm wide (Reprinted from [13] with permission from AAAS.)
Early microfluidic reactors were also primarily focused on biological applications. Aside from relatively simple reactions within microarrays, an important biological reaction that benefited from miniaturized formats is the polymerase chain reaction (PCR). PCR is a technique that uses enzymatic transcription to systematically amplify parts or all of a DNA strand, triggered via thermal fluctuations. On-chip PCR incorporation was made possible due to rapid and efficient heating and cooling of extremely small sample volumes, allowing for quick and proficient thermocycling. In 1995, Northrup et al. developed the first microfabricated device capable of thermocycling and PCR reactions [16]. On-chip PCR would prove to be instrumental in future DNA sequencing and genotyping devices.
1.2.4 Integrated systems Modular system integration is particularly advantageous in µTAS devices for several reasons. First, the sample can be isolated from the outside environment, reducing error caused from human contact and sample contamination. In addition, having all processes located on a single chip can, in theory, reduce sample-processing time and allow for a fully-automated analysis. This potential has prompted the idea of integrated devices for point-of-care diagnostics and clinical analysis. However, full integration of numerous modular components is not a trivial task and requires the resolution of numerous problems, including sample injection, pumping, data dissemination and product retrieval.
Introduction to Microfluidics
9
At the onset of modular device development, integrated systems were generally simple and relatively crude. Technological efforts were focused on designing simplistic systems that performed one or two actions instead of many. One of the first integrated DNA analysis systems was presented by Burns et al. in 1998 (Fig. 1.4). This device was engineered to meter a precise DNA sample volume, mix it with reagents, amplify it, separate it using gel electrophoresis, and detect the fluorescence signal on-chip [17]. Since that time, DNA separation stages have also been equipped with both on-chip fluorescent and conductivity detectors for on-chip analysis. However, a problem plaguing many of these systems, from a diagnostic standpoint, is that they lack an easy and reliable way to process raw sample, such as blood or saliva.
Fig. 1.4 Integrated DNA analysis device developed by Burns et al. in 1998 (Reprinted from [17] with permission from AAAS.).
During its infancy, microfluidics exhibited the potential to explode into a field rich with powerful applications and commercial successes. Some likened its promising impact to that of the integrated circuit (IC) industry. Today, however, the microfluidics field is arguably not where it was envisioned ten years ago. So what factors have impeded to the field’s emergence as a prevailing technology? We will further explore that question later on in the chapter, but for now we remind the reader that microfluidics continues to be a vibrant and active research area with researchers tackling the challenges hindering its widespread acceptance.
10
Sommer et. al
1.3 Fluidics and Transport Fundamentals Researchers employ many fundamental transport and scaling principles for manipulating and analyzing microfluidic flows. In this section we introduce some of the important concepts governing microscale transport, with the hope of elucidating several advantages of microfluidics over macroscale techniques. We also introduce several dimensionless groups (Table 1.2) that can be used to evaluate the relative importance of different phenomena.
1.3.1 The continuum approximation Unlike solids, fluids (especially gases) consist of molecules that are reasonably widely separated. However, in fluid mechanics, despite the fact that properties like velocity and density vary wildly at the molecular scale, we usually view fluids as “continuous” and discuss “average” fluid properties rather than considering the properties of each molecule. Does this continuum approximation still hold in microfluidics, where fluid volumes are very small? Indeed, for the volumes typically encountered in microfluidics, the continuum approximation remains sufficiently valid. A 1 picoliter [(10 µm)3] volume of fluid still contains 3 x 1013 water molecules, large enough for us to consider their average rather than individual behavior. Typically, for most properties the continuum approximation does not break down until we approach length scales on the order of several molecular diameters [18]. The continuum approximation is important because it allows us to analyze microfluidic flows with the same governing principles developed for macroscale fluid mechanics. However, as researchers continue to drive fluids into smaller length scales (i.e. nanofluidics and beyond), the continuum approximation starts to break down and we must develop new approaches for analyzing these small regimes.
1.3.2 Laminar flow Microfluidic devices almost always boast smooth laminar flow, as opposed to turbulent flow which is of a stochastic nature and marked by the presence of “eddies” that disrupt parallel streamlines. The Reynolds number
Introduction to Microfluidics
11
(Re) is a dimensionless parameter used to determine the transition from laminar to turbulent regimes, with Re < 2100 considered laminar for flow in cylindrical channels. The Reynolds number for this flow is defined as
Re =
ud
(1.1)
ν
where u is the flow velocity, ν is the kinematic viscosity, and d is the channel diameter. In fluid mechanics terms, the Reynolds number compares the magnitudes of inertial force to viscous forces in a flow. Because Re ∝ d , the small dimensions of microfluidic channels are responsible for very low Reynolds numbers, resulting in laminar flows. In fact, for most microfluidic applications Re < 1. Since microfluidic flows are usually laminar, simple flows like Poiseuille flow are commonly encountered. Poiseuille flow occurs when we have steady, fully-developed pressure driven flow of a Newtonian fluid in a channel. The velocity profile for Poiseuille flow is parabolic, with the maximum velocity being in the center of the channel. The equations for Poiseuille flow in a cylindrical channel are as follows [19]
r v z (r ) = 2U [1 − ( ) 2 ] R R 2 dP 8µ dz
(1.3)
πR 4 dP 8µ dz
(1.4)
U =−
Q=−
(1.2)
where v z is the velocity of the fluid along the axis, U is the mean velocity (maximum velocity at centerline of channel = 2U), R is the radius of the channel, r is the radial distance from the axis, z is the distance along the axis, P is the pressure along the channel, µ is the fluid viscosity and Q is the volumetric flow rate. While laminar flow is advantageous in many microfluidic applications (i.e. in electrophoretic separations for reducing band broadening due to dispersion, or in diffusion-based separation devices such as the H-filter), it can also be a nuisance, such as in processes where mixing is necessary. Most chemical and biological assays require mixing, or dilution, at some point.
12
Sommer et. al
Several clever techniques have been developed to achieve mixing in laminar regimes, mostly through geometric design or by taking advantage of enhanced diffusivity across small dimensions.
1.3.3 Diffusion in microfluidic systems Diffusion – the stochastic process by which molecules drift from one region to another – is another property that takes on increasing importance in microfluidic systems as channel dimensions are reduced. Diffusive transport is driven by random thermal motion of particles such that, given enough time and the absence of external influences, a species will be homogeneously distributed throughout a finite volume. Most microfluidic systems combine diffusive transport with convective flow of the bulk fluid; therefore it is helpful to compare the relative importance of each effect. The Sherwood number, a dimensionless number representing the ratio of convective mass transfer to diffusive mass transfer in a system, is defined as
Sh =
kd D
(1.5)
where k is the mass transfer coefficient, d is the characteristic length of the system (e.g. channel diameter), and D is the diffusion coefficient. For most macroscale systems, Sh is large, resulting in convective transport being dominant over diffusive transport. However, for microfluidic systems, the Sherwood number is much lower due to the presence of the characteristic length scale in the numerator. Therefore, diffusion assumes much more importance in microfluidic systems. Diffusive transport without flow (or perpendicular to streamlines with flow) can be estimated from the simplified mass transfer equation:
∂ 2C ∂C =D 2 ∂x ∂t
(1.6)
Consider a stationary liquid system with a step change in species concentration such that the concentration is held at C* for all time t at a certain location. Let there be no other source for the substance in the liquid and no other mode of transport besides diffusion. At time t, the concentration will be C*/2 at a distance of about
Dt from the source, and will be 1% of C*
Introduction to Microfluidics
13
at a distance of about 4 Dt (Fig. 1.5). Therefore, diffusion has a significant effect up to a distance of about 4 Dt . For polymers or proteins in a fluid, D is on the order of ~10-7 cm2/s. So for a time period of about 10 seconds, the distance at which diffusion is significant will be about 40µm, a length that is usually negligible for macroscale purposes but can be a significant distance in microfluidic systems. Note that, at twice that distance (i.e. 8 Dt ), the concentration is only 10-8C*.
Fig. 1.5 Diffusion-dependent concentration profile at time t for a solute with diffusion coefficient D.
Several mixing and separation techniques have been developed exploiting this diffusive effect in microfluidic systems. Readers are especially referred to the H-filter [20] developed by Paul Yager’s lab at the University of Washington for an example of an efficient and highly-effective device that uses this very simple concept (Fig. 1.6). Here, species in a stream are separated into two diverging flows based on their differing diffusivities across a microchannel.
14
Sommer et. al
Fig. 1.6 Conceptual representation of the H-filter developed in Paul Yager's lab at the University of Washington. Small solutes are filtered out of the sample stream based on their enhanced diffusivity transverse to the flow. Reprinted by permission from Macmillan Publishers Ltd: Nature [21], © 2006.
1.3.4 Surface forces and droplets Another important difference between fluid motion at the macroscale and the microscale is the enhanced significance of surface tension. The Bond number (Bo), defined as
Bo ≡
Δρ ⋅ a ⋅ L2
(1.7)
γ 12
represents the dimensionless ratio of body forces to surface tension forces at a fluid-fluid interface. Here Δρ is the density difference across the interface, a is the acceleration associated with the body force, which in most cases is gravity, L is the pertinent length scale (typically the radius of a droplet or diameter of a channel), and γ12 is the surface tension between the two fluid phases. The magnitude of the body forces relative to the surface tension forces rapidly decreases as the length scale is reduced, thereby reducing the Bond number. Fluidic control systems based on surface tension forces most commonly use changes in channel geometry or surface hydrophobicity to induce or inhibit pressure-driven flow. Surface tension forces can help drive fluid flow by exploiting capillary action in a phenomenon often described as wicking. In wicking, intermolecular forces between the fluid and a narrow channel’s surface trump those
Introduction to Microfluidics
15
within the fluid such that it is drawn through the channel. A simple application of this phenomenon is that of a blood viscometer, in which the sample’s viscosity is measured based on the distance it is drawn through a narrow channel [22]. Similarly, flow can be inhibited by altering the hydrophobicity of a channel surface. Several chemical and physical processes have been used to create hydrophobic stops at desired locations within microfluidic networks, such that the flow of an aqueous solution will be impeded upon reaching that location [23]. Surface tension is particularly significant in droplet-based - also referred to as digital - microfluidic systems due to the increased surface area-tovolume ratio that accompanies decreased fluid volumes. These digital systems most often utilize two immiscible phases with the continuous phase being a hydrophobic fluid, such as mineral oil or air, and the disperse phase being an aqueous solution, as is characteristic of biological samples. Droplets help suppress unwanted Taylor dispersion and evaporative effects, and have been used for many applications, including sample transport, mixing, and particle synthesis [24]. Electrowetting is a technique in which electric potentials are used to alter the local surface wettability of a substrate at opposing ends of a droplet, thereby inducing a pressure gradient within the liquid and subsequent droplet motion. Thermocapillary pumping is another example of droplet transport exploiting surface tension, in which one end of a droplet is heated to create a surface tension gradient and induced motion [25]. Table 1.2 Important dimensionless numbers in microfluidic systems Dimensionless Definition Significance Microfluidic advantage number Reynolds Ratio of inertial Typically Re is small: ud Re = Results in laminar flow for number forces to ν most microfluidic applica(Re) viscous forces tions. Sherwood number (Sh) Bond number (Bo)
Sh =
Bo ≡
kd D
Ratio of convective to diffusive mass transfer
Δρ ⋅ a ⋅ L2 Ratio of body
γ 12
forces to surface tension forces
Typically Sh is small: Diffusion is more important with smaller dimensions. Typically Bo is small: Enables pumping via capillary pressure and droplet-based transport systems.
16
Sommer et. al
1.3.5 Pumps and valves While the reduced dimensions of microfluidic systems create several advantages from a transport viewpoint, designing accurate pumping and valving control systems at this scale can be an arduous task. Nonetheless a wide variety of pumping methods for generating pressure-driven flows in microfluidic systems have been developed. The most straightforward method is direct application of externally controlled liquid sources such as syringe pumps, which are most suitable for systems containing high hydraulic resistances. Direct application of external pressure sources has also been demonstrated using both temporal and mechanical attenuation mechanisms to generate pressure differences appropriate for precise control of microscale flows. Several microfabricated pumping techniques have been developed with either off- or on-chip actuation. Serial deflection of PDMS membrane valves using external pneumatic control has gained widespread acceptance for generating pulsatile flows [26]. Similar PDMS valve designs substituting pneumatic actuation with external mechanical actuators, such as Braille pins, have also been reported [27]. Integrated actuation mechanisms, including electrostatics, piezoelectrics, electromagnetics, and thermal expansion, have been successfully implemented for actuation of a variety of other diaphragm materials. In addition to reciprocating pumps, microfabricated flow control components employing principles as diverse as thermal transpiration [28] and the Venturi effect have also been demonstrated [29].
1.3.6 Electrokinetics Electrokinetics encompasses a range of techniques for inducing motion of charged particles or conductive media by application of electric fields. Electroosmosis is the process by which the bulk fluid in a microchannel or porous media is “pumped” under application of an electric field. When an electrolyte solution is placed adjacent to a charged surface, such as glass, a layer of counter ions called the Stern layer adsorbs to the surface. Between the Stern layer and the bulk liquid is a region of intermediate charge density known as the diffuse electrical double layer. Movement of the double layer, and subsequently the bulk liquid, can be induced by electrostatic forces resulting from application of a DC potential along the channel axis. Electroosmotic flow results in a flat, plug-like velocity profile across the channel width, unlike the parabolic flow profile typical of pressure driven flow (Fig. 1.7).
Introduction to Microfluidics
17
Fig. 1.7 Physical representation of electroosmotic flow. Counterions (+) in bulk solution amass near the negatively charged surface in the electrical double layer. These ions migrate with the application of an electric field, and drag the remaining bulk solution via viscous forces. Electroosmosis results in a “flat” velocity profile, as opposed to the parabolic profile of pressure driven flow.
Electrophoresis, on the other hand, is simply motion of a charged particle in solution under an applied electric field. Most biological molecules and particles carry a surface charge that will induce a force in the direction of the electric field. The electrophoretic velocity, u, of a particle is governed by u = μE, where μ is the electrophoretic mobility of the particle (dependent primarily on its size and net charge) and E is the electric field strength. Electrophoresis allows for separation of different species (e.g. DNA, proteins, etc.) based on unique electrophoretic mobilities. Several electrophoretic techniques have been developed and implemented in microscale formats for enhanced separation and enrichment of charged analytes. For example in isoelectric focusing (IEF) [30-32], species migrate through an imposed pH gradient and drift toward their respective isoelectric point the pH at which the molecule’s net charge is zero. An analogous technique is temperature gradient focusing (TGF) [33, 34], in which species are separated by their temperature-dependent mobilities along an imposed temperature gradient. Another example includes isotachophoresis (ITP) [35], in which species are separated based on their different electrophoretic mobilities as they migrate between trailing (slow) and leading (fast) electrolytes. Droplets, as well as particles, suspended in a liquid medium can alternatively be manipulated using dielectrophoresis (DEP). The principle of DEP does not rely on the charge of particle but rather the relative polarizability of the particle with respect to the surrounding medium. In the presence of a non-uniform electric field, particles possessing a greater po-
18
Sommer et. al
larizability than the surrounding medium experience a force toward regions of higher field strength (positive DEP); conversely, particles less polarizable than the medium experience a force in the direction of lower field strength (negative DEP). Because the dielectrophoretic force is proportional to the gradient instead of the polarity of the electric field, both DC and AC fields can be used. AC and DC dielectrophoresis have been extensively used to trap bacteria and cells [36, 37].
1.3.7 Thermal management As is the case in most biological systems and assays, many of the processes performed on microfluidic devices are highly temperature sensitive. For example, PCR requires cyclical temperature fluctuations of 94°C, 54°C and 72°C, while most restriction endonucleases in restriction digest reactions are inactivated at temperatures above 65°C [38]. When inducing thermal fluctuations with microfabricated heating elements, it is important to ensure that the heating is in an isolated region and not adversely affecting other processes in the device. Additionally, many microfluidic devices also contain electronic components which can produce Joule heating, and therefore need to be isolated in a similar manner. Many thermal isolation techniques have been developed. One of the most common techniques is to use thermally resistant silicon barriers, such as cantilevers and bridges, to protect sensitive portions of the device [39-44]. Other materials with low thermal conductivity have been used, including quartz [45], ceramics [46] and porous silicon [47, 48]. Additional thermal isolation techniques include thermal conduits, silicon back dicing and silicon back etching [38]. Thermal conduits work by enhancing heat conduction at places where low temperatures are desired. Back-dicing and back etching involve physically isolating the heat generation regions from regions that are desired to be kept cool. Another way to thermally isolate certain regions is by selective cooling, a process in which microchannels are backed with heat sinks [49].
1.4 Device Fabrication The past few decades have generated a wealth of new materials and techniques for fabricating microfluidic devices. The myriad fabrication tools available to researchers today allows for simple processing, improved functionality, and rapid and inexpensive prototyping. Here we outline the
Introduction to Microfluidics
19
standard materials and processes used in constructing microfluidic devices; these concepts will be explored in further detail in a subsequent chapter.
1.4.1 Materials Early microfluidic systems predominantly used silicon as a substrate due to the wealth of existing fabrication techniques in microelectronic production. Researchers used standard photolithography and etching processes to construct microfluidic channels of precise dimensions on silicon wafers. As research progressed, however, the focus shifted from silicon to glass substrates. Glass presented a number of advantages over silicon. Perhaps its most obvious advantage is that glass is transparent, thereby allowing visualization of on-chip processes as well as simple detection for separation assays. Additionally, glass is more compatible with electroosmotic flow than conductive silicon, which was arguably the dominant pumping mechanism at the time. From a fabrication standpoint, glass can be bonded to a second substrate more easily than silicon. In general, although the cost of glass and silicon wafers is about the same, glass became regarded as a simpler and more universal substrate for microfluidics. For many applications, glass and silicon soon gave way to polydimethylsiloxane (PDMS), an elastomeric polymer used in everything from contact lenses to bathroom caulking. When cured, PDMS behaves like an elastic solid that maintains its molded structure. The use of PDMS for microfluidic applications was first presented by George M. Whitesides in the mid1990’s [50, 51]. In this method the elastomeric monomer is poured over a master mold structure (typically a silicon wafer with an inverse photoresist structure of the channel geometry) and then cured (Fig. 1.8). The cured structures can simply be peeled from the master and bonded to a planar substrate, forming the microchannel structures. The master mold can be reused multiple times to replicate devices. Thus, this method is especially desirable due to its low cost, flexibility, and rapid prototyping. Other polymers have also been used for microfluidic applications, including polymethylmethacrylate (PMMA) and polycarbonate (PC), and are typically fabricated using hot embossing and injection molding methods. PDMS, however, remains the dominant polymer used in microfluidic fabrication. Although PDMS gained widespread acceptance in the field, glass and silicon certainly still have utility in microfluidics. Polymers typically have inferior chemical resistance, aging, mechanical, thermal, and optical properties than glass and silicon. For instance, PDMS can restrict detection of
20
Sommer et. al
low wavelength fluorescence (near 400 nm). Thus for laser-induced fluorescence (LIF) detection, sensitivity is lower in PDMS than in comparative glass devices. Silicon has the advantage that it can combine on-chip electronics (i.e. via CMOS) with microfluidic networks. Also, silicon fabrication still supports the most precise geometrical tolerances, which will be increasingly important as channel dimensions continue to decrease. There is little doubt that we will see continued application of all three substrate materials (silicon, glass, and polymers) as microfluidic research proceeds.
Fig. 1.8 Schematic representation of PDMS casting using a silicon master mold with features fabricated using photoresist. Reprinted by permission from Macmillan Publishers Ltd: Nature [52], © 2007.
1.4.2 Fabrication and assembly Adopting the fabrication technology used in the microelectronics industry, microfluidic researchers frequently use lithographic techniques for creating the features needed in their device. Photolithography involves exposing a substrate coated with a photo-sensitive material, called a photoresist, to light such that the selectively developed regions can be shielded from (or subjected to) subsequent fabrication processes such as etching or deposition. Exposure transfers the 2-D pattern of desired features via a pho-
Introduction to Microfluidics
21
tomask placed between the substrate and light source. Material can then be removed from the substrate using etching techniques (wet or dry). Likewise certain materials (i.e. polymers and metals) can be embossed to the substrate surface using various deposition methods. Microfluidic channels are then typically enclosed by bonding the fabricated substrate to a planar matting material. This is usually accomplished using anodic bonding (glass-silicon), fusion bonding (silicon-silicon), or by depositing a thin film of polymer or adhesive between the substrates that will bond upon hardening. More traditional fabrication approaches have also been explored for constructing microfluidic devices. Conventional machining methods (i.e. milling, drilling, cutting, and turning) that remove material from metals and hard plastics have been used to create simple microstructures, primarily for molding applications [53, 54]. The benefit of these methods is that the mold can be constructed from robust materials like stainless steel that have long lifetimes. This method is generally only suitable for simple planar channel structures with dimensional tolerances on the order of tens of microns. Injection molding is by far the most popular method used in the production of polymer parts on the order of millimeters to centimeters or larger, and has been used to produce microfluidic devices [55, 56]. A polymer melt is injected under high pressure into an evacuated die cavity containing the desired master mold, and subsequently cooled and solidified such that it can be separated from the mold. While probably not practical for simple prototyping, injection molding would be a desirable fabrication tool for high throughput production of commercial devices. Threedimensional polymer microdevices have been fabricated using stereolithography, a bottom up fabrication process that uses a focused laser to optically cure polymer at precise coordinates in real time [57]. The advantage is that the fabrication is very simple in comparison to other techniques as it requires no masks; the structure is defined using a CAD program which allows for rapid and inexpensive prototype adjustments. However, this system is somewhat advanced and might be time consuming and cost prohibitive for the construction of simple microfluidic devices.
1.5 Biological Applications While microfluidics seemingly has application in many areas, such as manufacturing and aerospace disciplines, thus far it has almost exclusively been pursued in the biological sciences. Here we briefly explore some of
22
Sommer et. al
the most prevalent biological applications as a prelude to the more in-depth reviews presented in subsequent chapters of this text.
1.5.1 Genetic analysis (DNA/RNA) In the past few decades, genomic research has yielded powerful techniques with applications ranging from genotyping to clinical diagnosis. Completion of the human genome sequencing brought with it a high demand for rapid, high throughput genetic analysis. Microfabricated systems have become desirable solutions for rapid and inexpensive genetic assays that offer low sample consumption, low fabrication cost, and short operating time. Expanding on early developments in modular reaction and separation systems, integrated genetic assays have evolved into fully-functional microfluidic platforms. As the key component of most genetic assays, PCR has been studied extensively in microsystems. PCR microsystems can be categorized into two major groups: static flow-based systems, in which the reaction mixture remains in a chamber and is heated with external or on-chip heaters, and continuous flow-based systems, in which the reaction mixture flows through different temperature zones at a controlled flow rate [58]. Novel approaches, like thermal convection-based PCR, have also been reported [59]. Other genetic analysis techniques, including DNA sequencing, reverse transcript PCR, restriction digestion, ligase detection reaction (LDR), and capillary electrophoresis (CE) have also been successfully performed in microdevices. Recently, integrated microsystems that contain sample preparation, fluid handling, bioreaction, and product detection (either electrophoresis or microarray hybridization) have been constructed, providing the capability to quickly collect genetic information from raw biological samples [60]. In addition to genetic analysis, microfluidic devices combined with microelectronics have also been applied to investigate the dynamic behavior of single DNA or RNA molecules, and the interaction between DNA/RNA and proteins.
1.5.2 Proteomics In the post-genomic era, proteomics has become one of the most studied and challenging areas of biological interest. A map of the human proteome, similar to that of the genome, would provide enormous knowledge to biologists studying the effects of the environment, diseases, and drugs on the body [61-63]. Such a goal requires the development of rapid and
Introduction to Microfluidics
23
high throughput devices for molecular investigations of cellular compositions. Proteins are also of interest for clinical diagnostics, in which relative concentrations of protein biomarkers in physiological fluids can signal the presence of a systemic disease [64, 65]. Due to the diversity of amino acid chains in their composition, proteins are more complicated and show a much wider range of structures than DNA or RNA. Also unlike DNA and RNA, proteins have no self-amplification procedure; therefore dilute samples must be preconcentrated or enriched prior to analysis in order to overcome instrumental limits of detection. Moreover, proteins are highly unstable thermally and physically meaning careful design and operation is needed when manipulating proteins in microfluidic devices. Microfluidic protein analyses require sample pretreatment steps (i.e. extraction, dialysis, and enrichment) followed by the assay itself (i.e. immunoassays, enzymatic assays, or electrokinetic focusing and separations). While a great deal of research has been conducted on individual phases of proteomic analysis, full-integration is an arduous task and remains the objective of several efforts. With integration in mind, interfacing microfluidic devices with traditional identification methods may prove particularly advantageous for proteomic research. For example, mass spectrometry methods, such as MALDI-TOF and ESI-MS, are used frequently for characterizing unknown protein samples. Therefore to facilitate rapid mass analysis of proteins, many researchers are developing devices for interfacing microfluidics with mass spectrometers [66].
1.5.3 Cellular assays Due to the controllability and reproducibility of microfluidic systems, miniaturization of cell cultures and assays has been studied intensively. Microfluidics has been shown to provide more genuine in vitro environments than traditional cell culture techniques due to efficient heat and mass transport. Serial processing and parallelization have also enabled high throughput assays on a single chip [67]. For eukaryotic cells, biologically compatible materials (such as PDMS) can be structured as an extracellular matrix using the microfabrication methods mentioned earlier. This control of surface shape has been used to investigate cell physiology in various cultivation environments [68]. In addition, taking advantage of well-controlled laminar flow enables precise microfluidic cell treatment, such as chemical gradient or temperature steps.
24
Sommer et. al
These subcellular changes of cultivation environment are generally driven by diffusive mixing in a laminar flow [69, 70]. The combination of microfluidics and cell cultivation is also widely used in the field of microbial research. For example, a microfluidic chemostat for monitoring microbial growth was developed by S.R. Quake and his colleagues [71]. Bacteria immobilization, control of cell morphology, and dynamic bacterial population analyses have also been facilitated with microfabricated devices. Another important research area is that of single cell assays. Single cell assays eliminate the ensemble averaging effect of cell populations and enable the precise investigation of individual cells. A wide range of isolation methods have been developed, such as hydrodynamic focusing, microdroplet generation, microwell arrays, and physical barrier arrays [72]. Cell sorting and screening is also one of the most rigorously investigated topics. Cells can be screened and sorted from a continuous flow according to electrophoretic mobility, refractive index, or size. In particular, microscale flow assisted cell sorting (micro-FACS) systems implementing DEP or electroosmotic flow have been extensively pursued [73-75].
1.5.4 Drug delivery and compatibility In addition to their extensive applications in analysis and diagnosis, microfluidic devices also offer a host of benefits for in vivo applications, particularly for drug delivery. Microfluidic drug delivery devices not only deliver proper concentrations of functional drugs to certain target sites with controllable release rates and dosage, but also present some other advantages: reduced size and power consumption, simple operation, and the ability to achieve complex release patterns (e.g. continuous or pulsatile release). Drug delivery microsystems can be categorized into three main groups [76]: (a) biocapsules and microparticles for controlled and/or site-specific drug release, (b) microneedles for transdermal and intravenous delivery, and (c) implantable microsystems. Biocapsules usually contain micro/nano-porous biocompatible membranes for drug encapsulation. Today, microparticles with various size and shape can be produced using standard microfabrication techniques. Combined with proper coatings, these microparticles provide effective site-specific delivery, especially for peptide and protein-based drugs. Microneedles have been fabricated as both 1-D
Introduction to Microfluidics
25
in-plane arrays and 2-D out-of-plane arrays, providing transdermal drug delivery with minimal tissue damage and negligible pain sensation [77]. Most of the microneedle-based devices are also integrated with certain micropumping components to control the drug dosing. Implantable microsystems are more suitable for therapies that require many injections for a certain period of time, because they can not only reduce the injection times, but also precisely control the rate of drug release. With the development of implantable microsensors [78] comes the potential to build “smart” drug delivery device (Fig. 1.9). These devices will function without any human intervention by: (a) real-time monitoring of target physiological conditions in a patient’s body using microsensors, (b) converting these conditions into some detectable signal, and (c) controlling the drug release through microactuators based on the signal analysis [76]. The resultant closed-loop drug delivery can provide the patient with a selfregulated treatment regiment. One important issue that has to be considered in designing microsystems for biological applications is biocompatibility of the device substrates how to interface these man-made devices with the relevant biological in vitro or in vivo environments so as to ensure proper functionality for a desired period of time. Devices must be able to withstand attacks from bioreactions and/or bodily immune responses (for in vivo applications), and should not cause any unwanted reactions or inhibitions. There is no standard list of biocompatible materials due to the differences in specific biological applications, as well as other details such as fabrication processes and the implanted location and lifetime of the device. Biocompatibility is a difficult property to control (and define) and thus remains the subject of much research.
26
Sommer et. al
c
Fig. 1.9 Implantable drug release device in which the gold membrane anode covering in (a) is removed in (b) via the application of an electric field to initiate release from a reservoir. The plot in (c) shows the intermittent release rate of the device over several days. Reprinted by permission from Macmillan Publishers Ltd: Nature [79], © 1999.
1.6 The Future 1.6.1 Potential demand/market for microfluidic devices The microfluidic technological impact has long been anticipated, but use of lab-on-a-chip devices is still limited. Although a few LOC products are now commercially available for applications such as DNA analysis, pro-
Introduction to Microfluidics
27
tein crystallization, and performing simple chemical reactions, the field continues to search for a so-called “killer application” [80]. For healthcare alone, the potential market for simple and inexpensive diagnostics seems to be incredibly promising. In the future, healthcare providers will likely shift from treating diseases to anticipating and preventing them. Widespread screening or testing will be a necessity for such a system, and microfluidic platforms are the most plausible technology for successful implementation. Consider, for example, genetic screening of newborn children for indications of potential health risks. For a quick, orderof-magnitude approximation of the number of tests that might be required each year for such analysis in the United States alone, we can multiply the approximate number of genes times an estimated number of variants per gene [81], times the number of births each year [82] and arrive at 103 x 102 x 106 = 1011 tests per year. Of course this number represents a broad approximation, and the term “test” is not equivalent to “device” (i.e., one device can perform many tests), but the number does imply that there is potentially a large market for diagnostics that are as simple to acquire and operate as commercial home pregnancy tests.
1.6.2 Current products We need to remember that the field of microfluidics is still very much in its early adolescence. However, a number of companies are now developing and marketing microfluidic lab-on-a-chip systems for use in various areas such as biomedical research, environmental testing, and medical diagnostics. Micronics mass-produces disposable microfluidic ‘lab cards’ that can be used in the development of devices for various applications, most notably low-cost disposable cards to analyze disease in the field. Another established microfluidics-based company is Agilent, whose 2100 Bioanalyzer® is a successful device for analyzing DNA, RNA, proteins and cells using a microfluidic format. Caliper Life Sciences produces the LabChipTM system, which has become widely used by pharmaceutical firms for cellular and enzymatic screening. Other successful microfluidic products used in research labs and hospitals include the GeneXpert® system by Cepheid for amplifying and detecting target DNA, the Triage® BNP test by Biosite for assessing the severity of a heart failure, and the iSTAT® 1 system by Abbott for conducting rapid, comprehensive blood screening with minimal sample volumes. Microdevices for liquid dispensing have been developed in various formats by TTP Labtech, Perkin Elmer, Tecan, and Labcyte. These liquid dispensers can manipulate vol-
28
Sommer et. al
umes as low as 5 nl with high precision. Silex Microsystems in Sweden has been adopting MEMS-based techniques to produce microfluidic labon-a-chip devices for the life sciences market. HandyLab, based in Ann Arbor, MI, also manufactures devices for nucleic acid testing. Fluidigm produces lab cards for protein crystallization and multiplexed genotyping using their multilayer soft lithography technique (Fig. 1.10). Nanogen offers a microfluidics-based array platform for DNA hybridization research to help users selectively probe target DNA. Unlike companies that commercialize microsystems for specific applications, Micronit Microfluidics is a commercial supplier and fabricator of microfluidic devices for researchers worldwide. The Stanford Microfluidics Foundry and the Michigan Nanofabrication Facility are other microfluidic fabricators known for soft lithography. These commercial foundries produce customized microfluidic devices from various materials, tailored to customers’ demands.
Fig. 1.10 Fluidigm’s BiomarkTM 48.48 Dynamic Array device for multiplexed genotyping. Reproduced with permission from www.fluidigm.com.
1.6.3 Challenges and the future Existing microfluidics-based companies have opened the door to device commercialization, but it seems that they have just scratched the surface of the potential market. So, just what is holding the field back? Why haven’t we seen an explosive flux of new products into the market? The answer, as it is with most emerging technologies, appears to be cost. Microfluidic devices are relatively expensive to manufacture due to the equipment and processes typically required. Additionally, these devices are competing with conventional platforms with which most customers are accustomed.
Introduction to Microfluidics
29
Microfluidic technology must significantly outperform or cost less than current products in order for customers to justify the switch from conventional technology. The market will forever be “user-driven” rather than “technology-driven”. User-friendliness is another major hurdle to microfluidic device commercialization. Microfluidic systems must be easily operated by non-experts. User-friendliness involves many challenges such as system integration, interfacing, and packaging. While these issues remain hurdles toward commercialization, they still receive little research or funding support. Devices must have macroscopic inputs and should employ user-friendly diagnostic concepts such as the simple yes/no indicators for the presence of antigens, antibodies, viruses, or other biological targets used in commercial products (e.g., pregnancy tests). Running real-life fluid samples like blood and saliva through a microfluidic device, however, typically brings more problems than using a purified laboratory solution. Therefore researchers also need to design devices such that they are operable in real-world environments, rather than solely in the laboratory. The future of microfluidics holds enormous potential as researchers continue to bring the technology out of the laboratory and into our real life. The innate advantages of microfluidics are too hard to ignore; we will undoubtedly see microfluidic devices replace conventional techniques as the field is further explored. Commercial success will drive future research efforts, and the field will expand with influences not only in the biological arena, but other disciplines as well. The remainder of this text will further explore many exciting and diverse research areas associated with this budding technology.
References 1. Golay MJE (1957) Vapor Phase Chromatography and Telegrapher's Equation. Anal Chem 29:928-932. 2. vanDeemter JJ, Zuiderweg FJ, Klinkenberg A (1956) Longitudinal diffusion and resistance to mass transfer as causes of nonideality in chromatography. Chemical Engineering Science 5:271-289. 3. Petersen KE (1982) Silicon as a mechanical material. In: Proceedings of the IEEE 70:420-457. 4. Terry SC, Jerman JH, Angell JB (1979) A Gas Chromatographic Air Analyzer Fabricated on a Silicon Wafer. IEEE T Electron Dev 26:1880-1886.
30
Sommer et. al
5. Manz A, Graber N, Widmer HM (1990) Miniaturized total chemical analysis systems: A novel concept for chemical sensing. Sensors and Actuators B: Chemical 1:244-248. 6. Manz A, Miyahara Y, Miura J, Watanabe Y, Miyagi H, Sato K (1990) Design of an open-tubular column liquid chromatograph using silicon chip technology. Sensors and Actuators B: Chemical 1:249-255. 7. Manz A, Harrison DJ, Verpoorte EJ, Fettinger JC, Paulus A, Lüdi H, Widmer HM (1992) Planar chips technology for miniaturization and integration of separation techniques into monitoring systems : Capillary electrophoresis on a chip. Journal of Chromatography A 593:253-258. 8. de Mello A (2002) On-chip chromatography: the last twenty years. Lab Chip 2:48N-54N. 9. Reston RR, Kolesar ES (1994) Silicon-micromachined gas chromatography system used to separate and detect ammonia and nitrogen dioxide. I. Design, fabrication, and integration of the gas chromatography system. J Microelectromech Syst 3:134-146. 10. Mathies RA, Huang XC (1992) Capillary array electrophoresis: an approach to high-speed, high-throughput DNA sequencing. Nature 359:167-169. 11. Harrison DJ, Fluri K, Seiler K, Fan Z, Effenhauser C, Manz A (1993) Micromachining a Miniaturized Capillary Electrophoresis-Based Chemical Analysis System on a Chip. Science 261:895-897. 12. Woolley A, Mathies R (1994) Ultra-High-Speed DNA Fragment Separations Using Microfabricated Capillary Array Electrophoresis Chips. Proceedings of the National Academy of Sciences 91:11348-11352. 13. Fodor S, Read J, Pirrung M, Stryer L, Lu A, Solas D (1991) Light-directed, spatially addressable parallel chemical synthesis. Science 251:767-773. 14. Schena M, Shalon D, Davis R, Brown P (1995) Quantitative Monitoring of Gene Expression Patterns with a Complementary DNA Microarray. Science 270:467-470. 15. Schena M, Shalon D, Heller R, Chai A, Brown P, Davis R (1996) Parallel human genome analysis: Microarray-based expression monitoring of 1000 genes. Proceedings of the National Academy of Sciences 93:10614-10619. 16. Northrup MA, Gonzalez C, Hadley D, Hills RF, Landre P, Lehew S, Saw R, Sninsky JJ, Watson R (1995) A Mems-based Miniature DNA Analysis System. In: Solid-State Sensors and Actuators and Eurosensors IX 1:764-767. 17. Burns M, Johnson B, Brahmasandra S, Handique K, Webster J, Krishnan M, Sammarco T, Man P, Jones D, Heldsinger D, Mastrangelo C, Burke D (1998) An Integrated Nanoliter DNA Analysis Device. Science 282:484-487. 18. Prasanth PS, Kakkassery JK (2006) Direct Simulation Monte Carlo (DSMC): A numerical method for transition-regime flows - A review. J Indian Inst Sci 86:169-192. 19. Batchelor GK (2000) An Introduction to Fluid Dynamics. Cambridge University Press, Cambridge, UK. 20. Brody JP, Osborn TD, Forster FK , Yager P (1996) A planar microfabricated fluid filter. Sensors and Actuators A: Physical 54:704-708.
Introduction to Microfluidics
31
21. Yager P, Edwards T, Fu E, Helton K, Nelson K, Tam MR, Weigl BH (2006) Microfluidic diagnostic technologies for global public health. Nature 442:412418. 22. Srivastava N, Davenport RD, Burns MA (2005) Nanoliter Viscometer for Analyzing Blood Plasma and Other Liquid Samples. Anal Chem 77:383-392. 23. Handique K, Gogoi BP, Burke DT, Mastrangelo CH, Burns MA (1997) Microfluidic flow control using selective hydrophobic patterning. Proc SPIE Conference on Micromachined Devices and Components 3224:185-195. 24. Shui L, Eijkel JCT, van den Berg A (2007) Multiphase flow in microfluidic systems – Control and applications of droplets and interfaces. Advances in Colloid and Interface Science 133:35-49. 25. Sammarco TS, Burns MA (1999) Thermocapillary pumping of discrete drops in microfabricated analysis devices. AICHE J 45:350-366. 26. Unger MA, Chou H, Thorsen T, Scherer A, Quake SR (2000) Monolithic Microfabricated Valves and Pumps by Multilayer Soft Lithography. Science 288:113-116. 27. Gu W, Zhu X, Futai N, Cho BS, Takayama S (2004) Computerized microfluidic cell culture using elastomeric channels and Braille displays. Proceedings of the National Academy of Sciences 101:15861-15866. 28. Namasivayam V, Larson RG, Burke DT, Burns MA (2003) Transpirationbased micropump for delivering continuous ultra-low flow rates. J Micromech Microengineering 13:261-271. 29. Chang DS, Langelier SM, Burns MA (2007) An electronic Venturi-based pressure microregulator. Lab on a Chip 7:1791-1799. 30. Sommer GJ, Singh AK, Hatch AV (2008) On-Chip Isoelectric Focusing Using Photopolymerized Immobilized pH Gradients. Anal Chem 80:3327-3333. 31. Wu X, Sze NS, Pawliszyn J (2001) Miniaturization of capillary isoelectric focusing. Electrophoresis 22:3968-3971. 32. Herr AE, Molho JI, Drouvalakis KA, Mikkelsen JC, Utz PJ, Santiago JG, Kenny TW (2003) On-Chip Coupling of Isoelectric Focusing and Free Solution Electrophoresis for Multidimensional Separations. Anal Chem 75:11801187. 33. Ross D, Locascio LE (2002) Microfluidic Temperature Gradient Focusing. Anal Chem 74:2556-2564. 34. Kim SM, Sommer GJ, Burns MA, Hasselbrink EF (2006) Low-Power Concentration and Separation Using Temperature Gradient Focusing via Joule Heating. Anal Chem 78:8028-8035. 35. Gebauer P, Bocek P (2002) Recent progress in capillary isotachophoresis. Electrophoresis 23:3858-3864. 36. Gagnon Z, Chang H (2005) Aligning fast alternating current electroosmotic flow fields and characteristic frequencies with dielectrophoretic traps to achieve rapid bacteria detection. Electrophoresis 26:3725-3737. 37. Cummings EB, Singh AK (2003) Dielectrophoresis in Microchips Containing Arrays of Insulating Posts: Theoretical and Experimental Results. Anal Chem 75:4724-4731.
32
Sommer et. al
38. Yang M, Pal R, Burns MA (2005) Cost-effective thermal isolation techniques for use on microfabricated DNA amplification and analysis devices. J Micromech Microengineering 15:221-230. 39. Kajiyama T, Miyahara Y, Kricka L, Wilding P, Graves D, Surrey S, Fortina P (2003) Genotyping on a Thermal Gradient DNA Chip. Genome Res 13:467475. 40. Hung S, Wong S, Fang W (2000) The development and application of microthermal sensors with a mesh-membrane supporting structure. Sensors and Actuators A: Physical 84:70-75. 41. Socher E, Bochobza-Degani O, Nemirovsky Y (2001) A novel spiral CMOS compatible micromachined thermoelectric IR microsensor. J Micromech Microengineering 11:574-576. 42. Ji-song H, Kadowaki T, Sato K, Shikida M (2002) Fabrication of thermalisolation structure for microheater elements applicable to fingerprint sensors. Sensors and Actuators A: Physical 100:114-122. 43. Mo Y, Okawa Y, Inoue K, Natukawa K (2002) Low-voltage and low-power optimization of micro-heater and its on-chip drive circuitry for gas sensor array. Sensors and Actuators A: Physical 100:94-101. 44. Lee K, Lee H, Byun H, Cho I, Bu J, Yoon E (2001) An audio frequency filter application of micromachined thermally-isolated diaphragm structures. Sensors and Actuators A: Physical 89:49-55. 45. Xie B, Mecklenburg M, Danielsson B, Öhman O, Winquist F (1994) Microbiosensor based on an integrated thermopile. Analytica Chimica Acta 299:165-170. 46. Chou CF, Changrani R, Roberts P, Sadler D, Burdon J, Zenhausern F, Lin S, Mulholland A, Swami N, Terbrueggen R (2002) A miniaturized cyclic PCR device—modeling and experiments. Microelectronic Engineering 61-62:921925. 47. Lysenko V, Périchon S, Remaki B, Barbier D (2002) Thermal isolation in microsystems with porous silicon. Sensors and Actuators A: Physical 99:13-24. 48. Kaltsas G, Nassiopoulou AG (1999) Novel C-MOS compatible monolithic silicon gas flow sensor with porous silicon thermal isolation. Sensors and Actuators A: Physical 76:133-138. 49. Cassia L, Di Pietro DA, Marengo M (2002) Micro-heat-exchangers: technical characteristics, production and market. Microscale Heat Transfer 1-12. 50. Xia Y, Kim E, Zhao XM, Rogers JA, Prentiss M, Whitesides GM (1996) Complex Optical Surfaces Formed by Replica Molding Against Elastomeric Masters. Science 273:347-349. 51. Kim E, Xia Y, Whitesides GM (1995) Polymer microstructures formed by moulding in capillaries. Nature 376:581-584. 52. Weibel DB, DiLuzio WR, Whitesides GM (2007) Microfabrication meets microbiology. Nat Rev Micro 5:209-218. 53. Martynova L, Locascio LE, Gaitan M, Kramer GW, Christensen RG, MacCrehan WA (1997) Fabrication of Plastic Microfluid Channels by Imprinting Methods. Anal Chem 69:4783-4789.
Introduction to Microfluidics
33
54. Locascio LE, Gaitan M, Hong J, Eldefrawi M (1998) In: Proc Micro-TAS '98 367-370. 55. McCormick RM, Nelson RJ, Alonso-Amigo MG, Benvegnu DJ, Hooper HH (1997) Microchannel Electrophoretic Separations of DNA in InjectionMolded Plastic Substrates. Anal Chem 69:2626-2630. 56. Paulus A, Williams SJ, Sassi SJ, Kao PH, Tan H, Hooper HH (1998) Integrated capillary electrophoresis using glass and plastic chips for multiplexed DNA analysis. In: Proc. Of SPIE 3515:94-103. 57. Ikuta K, Maruo S, Kojima S (1998) New micro stereo lithography for freely movable 3D micro structure-super IH process with submicron resolution. In: Proceedings of MEMS 98 290-295. 58. Senturia SD (2001) Microsystem Design. Kluwer Academic Publishers, Boston. 59. Krishnan M, Ugaz V, Burns M (2002) PCR in a Rayleigh-Benard Convection Cell. Science 298:793. 60. Pal R, Yang M, Lin R, Johnson BN, Srivastava N, Razzacki SZ, Chomistek KJ, Heldsinger DC, Haque RM, Ugaz VM, Thwar PK, Chen Z, Alfano K, Yim MB, Krishnan M, Fuller AO, Larson RG, Burke DT, Burns MA (2005) An integrated microfluidic device for influenza and other genetic analyses. Lab on a Chip 5:1024-1032. 61. Sanders GW, Manz A (2000) Chip-based microsystems for genomic and proteomic analysis. Trends in Analytical Chemistry 19:364-378. 62. Lion N, Reymond F, Girault HH, Rossier JS (2004) Why the move to microfluidics for protein analysis? Current Opinion in Biotechnology 15:31-37. 63. Freire SLS, Wheeler AR (2006) Proteome-on-a-chip: Mirage, or on the horizon? Lab on a Chip 6:1415-1423. 64. Srinivasan V, Pamula VK, Fair RB (2004) An integrated digital microfluidic lab-on-a-chip for clinical diagnostics on human physiological fluids. Lab on a Chip 4:310-315. 65. Herr AE, Hatch AV, Throckmorton DJ, Tran HM, Brennan JS, Giannobile WV, Singh AK (2007) Microfluidic immunoassays as rapid saliva-based clinical diagnostics. PNAS 104:5268-5273. 66. Koster S, Verpoorte E (2007) A decade of microfluidic analysis coupled with electrospray mass spectrometry: An overview. Lab on a Chip 7:1394-1412. 67. El-Ali J, Sorger PK, Jensen KF (2006) Cells on chips. Nature 442:403-411. 68. Balaban NQ, Merrin J, Chait R, Kowalik L, Leibler S (2004) Bacterial Persistence as a Phenotypic Switch. Science 305:1622-1625. 69. Takayama S, Ostuni E, LeDuc P, Naruse K, Ingber DE, Whitesides GM (2001) Laminar flows: Subcellular positioning of small molecules. Nature 411:1016-1016. 70. Tourovskaia A, FigueroaMasot X, Folch A (2005) Differentiation-on-a-chip: A microfluidic platform for long-term cell culture studies. Lab on a Chip 5:14-19. 71. Balagadde FK, You L, Hansen CL, Arnold FH, Quake SR (2005) Long-Term Monitoring of Bacteria Undergoing Programmed Population Control in a Microchemostat. Science 309:137-140.
34
Sommer et. al
72. Roman G, Chen Y, Viberg P, Culbertson A, Culbertson C (2007) Single-cell manipulation and analysis using microfluidic devices. Analytical and Bioanalytical Chemistry 387:9-12. 73. Hunt TP, Issadore D, Westervelt RM (2008) Integrated circuit/microfluidic chip to programmably trap and move cells and droplets with dielectrophoresis. Lab Chip 8:81-87. 74. Ahn K, Kerbage C, Hunt TP, Westervelt RM, Link DR, Weitz DA (2006) Dielectrophoretic manipulation of drops for high-speed microfluidic sorting devices. Appl Phys Lett 88:024104. 75. Li Y, Dalton C, Crabtree HJ, Nilsson G, Kaler KVIS (2007) Continuous dielectrophoretic cell separation microfluidic device. Lab Chip 7:239-248. 76. Razzacki ZS, Thwar PK, Yang M, Ugaz VM, Burns MA (2004) Integrated microsystems for controlled drug delivery. Advanced Drug Delivery Reviews 56:185-198. 77. McAllister D, Wang P, Davis S, Park J, Canatella P, Allen M, Prausnitz M (2003) Microfabricated needles for transdermal delivery of macromolecules and nanoparticles: Fabrication methods and transport studies. Proceedings of the National Academy of Sciences 100:13755-13760. 78. Receveur RAM, Lindemans FW, Rooij NFd (2007) Microsystem technologies for implantable applications. J Micromech Microengineering 17:R50-R80. 79. Santini Jr. JT, Cima MJ, Langer R (1999) A controlled-release microchip. Nature 397:335-338. 80. Blow N (2007) Microfluidics: in search of a killer application. Nat Meth 4:665-670. 81. Ladiges W, Kemp C, Packenham J and Velazquez J (2004) Human gene variation: from SNPs to phenotypes. Mutation Research 545: 131-139 82. "U.S. Census Bureau" (2008) accessed 2008. online. www.census.gov.
Chapter 2 Materials and Microfabrication Processes for Microfluidic Devices
Arash Noori, Sarvesh Upadhyaya, and P. Ravi Selvanganapathy McMaster University, Hamilton ON L8S 4L7, CANADA Correspondence should be addressed to: P. Ravi Selvanganapathy (
[email protected])
Keywords: Microfluidics, microfabrication, Silicon, Glass, Parylene, PDMS, Hydrogel, Paraffin, Polyimide, bulk micromachining, surface micromachining, micro injection molding, hot embossing, micro stereolithography
Abstract This chapter elaborates on the varied materials and microfabrication techniques used in the manufacture of microfluidic devices. The origins of these techniques traces back to semiconductor and precision machining industries. This chapter details microfabrication processes such as bulk and surface micromachining for silicon, glass and polymeric materials. It also discusses precision machining techniques such as micro injection molding, hot embossing, micro stereolithography. Over the past few decades, these techniques have been modified to suit the complexity and precision required in the 3-D construction of microfluidic devices. New materials such as parylene, poly-dimethylsiloxane, paraffin and hydrogels have been in-
36
Noori, Upadhyaya, and Selvaganapathy
troduced and microfabrication techniques developed that enhances the functionality of microfluidic devices. This chapter provides a comprehensive overview of material properties, microfabrication procedures and examples of its application in microfluidic devices.
2.1 Introduction Microfabrication technology originated in the semiconductor industry in response to the need to manufacture integrated circuits. These techniques were primarily lithography based and were developed for silicon based materials [1]. Since integrated circuits have their critical elements (transistors) in microscale dimensions and are packed tightly together, the fabrication technologies developed were targeted at achieving precision in the sub micrometer scale. Simultaneously, there has been a significant effort in the precision machining industry to develop materials and processes that achieve microscale precision for applications in a number of areas including micromotors, sensors, micro-optics and other precision components. A number of materials including metals, polymers and ceramics and process such as electrodischarge machining, precision milling, single point diamond turning, and laser micromachining have been developed over the years [2]. Several of these methods and processes for manufacture have been adopted and combined over the past 15 years to create microfluidic structures for applications in areas as diverse as medical diagnostics, drug delivery, drug discovery, analytical chemistry, combinatorial synthesis, molecular diagnosis, and specialty chemical manufacturing [3]. Laboratory based analytical methods used in areas of medical diagnostics and drug discovery are the gold standard in accuracy but are slow, have low throughput, and have high cost per analysis. Similarly, yields in special chemical manufacture are low due to the lack of micro level control over unit operations and residence times. Microfluidic devices offer the promise of miniaturizing and integrating multiple laboratory unit operations within a single chip. Miniaturized Lab-on-Chip (LoC) devices consume small volumes of samples and expensive reagents and can be made portable and available at the point of care. They also reduce the time of analysis significantly as unit operations like heating, mixing, and metering can be performed more accurately and quickly in smaller volumes thus providing real-time analysis. In the case of microreactors, they allow precise definition of process parameters such as interaction time between reactants by the structural design of the reactor, enabling high yields and fea-
Materials and Microfabrication Processes
37
sibility of certain reactions that were not possible before. Using microfabrication, several devices can be produced in parallel with the same processing steps, reducing their unit cost. Furthermore, the structural features and hence the functioning of the devices are accurate and repeatable which leads to accurate analysis and higher degree of confidence [4]. Several materials have been used in the construction of microfabricated devices such as silicon, glass, polymers, metals and ceramics. The most widely used material has been silicon, primarily due to the fact that early researchers were more familiar with the material and the processes involved in structuring it to the necessary shape. Glass is also used extensively in microfluidic devices, especially those targeted for biological applications, as most of the biochemical reactions have been characterized in glass. Lately, polymers have been incorporated into microfluidic devices due to the variety of surface properties obtainable and their ease of fabrication. This chapter deals with some of the widely used materials in microfluidic applications along with their processing techniques.
2.2 Silicon Based Materials The first miniature microfluidic system was a gas chromatographic system, microfabricated in silicon substrate, which was able to separate and analyze simple gas mixtures [5]. The system consisted of a sample injection valve and a 1.5m long separating capillary column fabricated on a silicon substrate and a nickel film as the sensing element of the detector. Silicon was the preferred material for initial microfluidic systems due to the availability of fabrication techniques and equipment from the microelectronics industry. Around the same time, research was ongoing at IBM for the development of inkjet printer heads [6, 7, 8] in silicon using anisotropic etching techniques. These developments can be seen as the origins of microfluidics. Over the years, with the development of more sophisticated fabrication techniques, a number of these microfluidic components have been integrated on a single substrate to automate and perform a complex series of unit operations for biochemical analysis [9]. Applications include DNA analysis [10], gas chromatography [11], chemical microreactors [12], forensic analysis [13], environmental analysis [14], and drug discovery [3]. Micro-total-analysis-systems offer a number of potential advantages when compared to conventional systems, including lower fabrication costs due to massively parallel fabrication techniques, reduced reagent consumption and dead-volume, and improved performance due to the possibility of per-
38
Noori, Upadhyaya, and Selvaganapathy
forming multiple analyses on a single chip [15]. Although other materials such as polymers are gaining in popularity, silicon continues to play a crucial role in the fabrication of microfluidic devices. A number of companies such as MEMSCAP, Colibrys, Intellisense, Integrated Sensing Systems Inc. (ISSYS), Innovative Micro Technology (IMT), Micralyne Inc., Micronit BV, Silex Microsystems and GE sensing offer foundry services or produce limited volumes of several microfluidic devices in silicon. Application areas include microneedles for drug delivery; devices for cell sorting, filtration and purification; DNA sensing and amplification devices; and sensing devices for flow, pressure, temperature, pH, and humidity. Some of the more commonly used materials are silicon, silicon-di-oxide, and silicon-nitride. Their material properties are shown in Table 2.1. Table 2.1 Material. properties of commonly used silicon based materials in microfluidics Properties Electrical Dielectric strength Resistivity Dielectric Constant
Silicon
Silicon Oxide
Silicon Nitride
1000ohm-cm 11.8
107 V/m 1016 ohm-m 3.9-4.3
107 V/m 1014 ohm-cm 7.5
Mechanical Tensile strength Young’s modulus Poisson’s ratio Elongation to break Density Hardness
2 GPa 160 GPa 0.27 1.2% 2.33 g/cm3 11.9-13 GPa
75 GPa 0.17 2.19 g/cm3 9 GPa vickers
260-330 GPa 0.23 - 0.27 3.1 g/cm3 16-18 GPa vickers
Thermal Thermal Conductivity Melting Temperature Coefficient of thermal expansion Specific Heat Capacity
150 W/m.K 1414 C
1.4W/m.K 1610 C
25-36 W/m.K 1900 C
2.49 x10-6 /C 0.702 J/g.K
0.5 x10-6 /C 1.0 J/g.K
2.8-3.2 x10-6 /C 0.54-0.7J/g.K
Other Properties: Index of refraction
3.42
1.46
2.05
Materials and Microfabrication Processes
39
2.2.1 Micromachining of silicon Silicon micromachining can be classified into two categories, bulk micromachining and surface micromachining. Of these two technologies, bulk micromachining is the most widely used and most commercially successful technology [16]. It involves the selective removal of the substrate material, in the regions unprotected by a masking layer, which allows the definition of features in the substrate. The masking layer itself is patterned using photolithographic process. Bulk micromachining has found applications in the manufacture of various devices such as pressure sensors and ink-jet print heads [17]. Conversely, surface micromachining is used to create microstructures on top of the substrate, rather than in the substrate itself, using photolithographic process. It involves the deposition of thin films on a patterned sacrificial layer, which is then etched away, leaving the thin film as the structural material. Lately, these surface micromachined techniques have gained importance in fabrication of microfluidic networks and components. 2.2.2 Bulk micromachining Micromachining processes that etch into the silicon substrate for device fabrication are referred to as bulk micromachining techniques. Bulk micromachining of silicon combined with other processes (such as metal deposition to create electrodes, chemical mechanical polishing, and subsequent wafer bonding) is used for the fabrication of microfluidic devices. Bulk micromachining processes can be further divided into etching using liquid etchants (wet etching) and etching using gaseous etchants (dry etching) [18]. The etch profile in these processes can be isotropic or anisotropic. Isotropic etchants etch uniformly in all directions, whereas anisotropic etchants are direction dependent and etch some directions preferentially over others [19]. As a result of this, isotropic etches result in rounded cross sections, whereas anisotropic etches result vertical sidewalls or those constrained by the crystal planes of the silicon substrate. This is shown in Fig. 2.1. Etch rates are characterized and etching process carried out for the required duration of time to obtain the desired feature dimension. Choice of etchant, masking material, and etch stop layer are all critical in the design of the bulk micromachining process [18]. Structures that have been fabricated by bulk micromachining include micro channels, micro wells, fluidic interconnects, micropumps, valves, mixers and reaction chambers for microfluidic applications.
40
Noori, Upadhyaya, and Selvaganapathy
Fig. 2.1 Cross sectional profiles of different etch methods: a.) Isotropic wet etch, b.) Anisotropic wet etch, c.) Isotropic dry etch, d.) Anisotropic dry etch.
2.2.2.1 Wet isotropic etching
Wet etching of silicon is widely used because of its fast etch times, low cost, low-complexity, and the availability of masking materials [20]. A commonly used isotropic wet etchant is a mixture of hydrofluoric (HF), nitric (HNO3), and acetic (CH3COOH) acids also known by the acronym HNA. When a silicon wafer is exposed to the etchant, it is oxidized by the nitric acid. The product is then etched by the hydrofluoric acid [19]. The details of this etch can be summarized into the following four basic steps. The process begins with the injection of electron holes into the Si to form higher oxidation state Si2+ or Si+. Hydroxyl groups (OH-) then attach to the positively charged Si, and initiate the reaction of the hydrated Si with the complexing agent in the solution. Finally, the reacted products are dissolved into the etchant solution. Therefore, any etching solution must provide a source of electron holes, hydroxyl groups, and a complexing agent whose reacted species are soluble in the etchant solution [18]. Depending on the ratio of the acids in the mixture, etch rates of silicon in HNA can vary from 0.1 to over 100 µm/min. Uniformity of etches is typically difficult to control but improves with stirring [19]. 2.2.2.2 Wet anisotropic etching
Anisotropic etchants such as potassium hydroxide (KOH), sodium hydroxide (NaOH) and Ethylenediamine - pyrocatechol (EDP) have etch rates dependent on the orientation of the crystallographic planes on the silicon wafer. On a silicon wafer, the [111] planes are at an angle of 54.74 degrees
Materials and Microfabrication Processes
41
for [100] wafers and 90 degrees for [110] wafers. In the anisotropic etchants, the etch rate along [111] plane is generally the slowest by 2-3 orders of magnitude. This property can be exploited with careful design to produce anisotropic structures and smooth sidewalls. For instance, at 80oC a mixture of potassium hydroxide (KOH) in water and isopropyl alcohol has an etch rate of about 2.1µm/min in the [110] plane, 1.4um/min in the [100] plane, but only 0.003µm/min in the [111] plane [16]. The higher etch rates in the [100] and [110] directions result in the exposure of the slow etching [111] planes over time. Figure 2.1 shows the orientation-dependent etching of a [100] silicon wafer with a patterned silicon dioxide mask. Using anisotropic etching, it is possible to accurately predetermine the result if the characteristics of the etchant are known [21]. With a proper mask design and orientation, one can obtain perfectly aligned square or rectangular cavities and pits, V-grooves, as well as holes or channels with vertical sidewalls [18]. Some of the earliest microfluidic devices were fabricated using anisotropic etching. Bassous et. al. [7] fabricated arrays of nozzles with square orifices for ink-jet printing applications on [100] wafers using an Ethylenediamine - pyrocatechol (EDP) etching solution. The nozzles consist of truncated square pyramidal cavities bounded by the four convergent [111] planes and 2 [100] surface planes as shown in Figure. 2.2. The anisotropic etches provide a very high degree of symmetry as a result of the crystallographic perfection of the starting wafer and a well controlled etch rate. The fabrication is carried out using conventional Si-processing techniques. The wafer is chemical-mechanically polished, cleaned, and thermally oxidized to form a 0.5µm thick SiO2 film. Photoresist is applied to both surfaces and patterned. Subsequently, the square openings are defined and etched in the SiO2 film to reveal the silicon wafer. The wafers are then etched in a solution of EDP, until the openings appear on the back side of the wafers. Finally, the wafers are cleaned, the oxide is stripped and the wafers are oxidized to provide a uniform oxide coating throughout the structure.
42
Noori, Upadhyaya, and Selvaganapathy
Fig. 2.2 Anisotropically etched pyramidical inkjet nozzle in silicon (Reproduced from [7] with permission of ECS – The Electrochemical Society).
In the previous example, the etching stopped when the orifices appeared on the back side of the wafer. Another method involves the use of timed etches, which may be used when tolerances for the final structure are not important or when the etch rate is well defined and reproducible [20]. However, when accuracy and tolerances are crucial, doping induced etch rate variations can be used as an etch stop. Highly boron doped silicon has a greatly reduced etch rate in all alkaline etchants [18]. By selective boron (p++) doping through a mask, regions can be made resistant to etching, while undoped regions will be etched away. For boron doping concentrations of ≥20x 1020cm-3, the KOH silicon etch rate of 0.7-3.0 µm is reduced by a factor of 20 [18]. This allows for accurate control of vertical dimensions, allowing the fabrication of membranes and other complex threedimensional microstructures [18]. The primary advantages of the ion implanted etch-stop techniques are independence of crystal orientation, smooth surface finish, and the ability to fabricate complex microstructures [20]. The main limitation is the maximum depth of ~15µm practically that is achievable using dopant diffusion [22]. Another method for the fabrication of microfluidic channels - developed by Chen et al [23] - incorporates doping, anisotropic etching, and isotropic etching for the fabrication of buried microchannels in silicon. The process flow is shown in Figure 2.3. The fabrication process begins with a p-type [100] silicon wafer with a patterned oxide layer that undergoes shallow boron diffusion to form a 2µm highly boron-doped layer on the surface. Using reactive ion etching (RIE), a chevron pattern is opened in the boron layer to define the channel structure. The features of the chevron structure are narrow enough to be undercut quickly in the subsequent wet etch, while preserving silicon in the bridge. To open up the channel structure, an
Materials and Microfabrication Processes
43
anisotropic silicon etch (EDP) is used to undercut this chevron structure as shown in Figure 2.4. The sidewalls of the flow channel are [111] surfaces, which have the slowest etch rate in EDP. Therefore, the width of the flow channel is determined by the area defined by the patterned chevron bridges instead of lateral etching. Although chevron patterns are used in this case as openings, other shapes could be used as well. After the channel is etched in the silicon, deep boron diffusion is performed to define the probe shank. Sealing of the channel is accomplished using thermal oxidation of a 0.5µm of thermal oxide and LPCVD dielectrics. Figure 2.4 shows the cross section of the completed sealed channel.
Fig. 2.3 Fabrication process for chemical delivery probe using undercutting (Reproduced with permission from [23] ©[1997] IEEE)
44
Noori, Upadhyaya, and Selvaganapathy
Fig. 2.4 Cross section of chemical delivery probe fabricated using chevron pattern and undercutting (Reproduced with permission from [23] ©[1997] IEEE).
2.2.2.3 Dry etching - Deep Reactive Ion Etching
Dry etching of silicon can be achieved using Deep Reactive Ion Etching (DRIE). In this technique ions and other chemically reactive species generated in low pressure plasma are accelerated towards the substrate producing anisotropy and enhancing reactivity. DRIE allows for the fabrication of very high-aspect-ratio structures with practical etch depths of 1mm and etch rates in the order of 2-3 µm/min [22]. A variation of DRIE, which includes alternate etching and side wall passivation steps to enhance anisotropy, was developed by Bosch and is generally referred to as the Bosch process. The process relies on an inductively coupled plasma source and an alternating etching and deposition cycle [24]. During the etch cycle, the wafer is exposed to SF6 to etch the exposed silicon. A protective polymer film consisting of CF2 molecules is then blanket deposited using a C4F8 gas source. This is followed by another etch in which the protective polymer at the bottom of the trench is removed by SF+ ions, without damaging the protective film on the sidewalls [20]. This process is repeated to obtain very high-aspect-ratio structures with anisotropies in the order of 30:1 and sidewall angles of 90+/-2 degree. Furthermore, it provides polymer and silicon dioxide selectivities of 50-100:1 and 120-200:1, respectively [22]. The DRIE process has been used extensively for the fabrication of through holes and deep microchannels in microfluidic devices. Li et al. [25] used DRIE in their fabrication of a microfluidic gas centrifuge for the separation of dilute gas mixtures. Figure 2.5 shows the fabrication process for the device. First, a 2µm layer of oxide is grown on the backside of a silicon wafer to act as an etch mask. After this, a 1.5µm layer of photoresist is deposited and patterned on the front side of the wafer, and 5µm deep gas concentrating channels are etched using reactive ion etching. A 6µm layer of photoresist is then spun and patterned for the inlet and outlet ports, after
Materials and Microfabrication Processes
45
which the oxide is etched with buffered HF. These unprotected openings serve as the locations where the DRIE will take place. After the DRIE is performed, the resist and oxide are stripped, and the silicon wafer is anodically bonded to a Pyrex wafer to seal the channels. As a final step, metal capillaries are inserted into the inlet and outlet ports in the silicon wafer to serve as the fluidic interfaces. DRIE was used in the fabrication of this device as existing fabrication methods of isotope separation devices relied on elaborate and costly process of stacking of photo-etched metal foils or micromachining using a combination of lithography, electroplating, and injection molding known by its German acronym LIGA.
Fig. 2.5 Fabrication process of microfluidic device using DRIE for the creation of interconnect holes. Reprinted from [25] with permission from Elsevier.
46
Noori, Upadhyaya, and Selvaganapathy
2.2.3 Surface micromachining Surface micromachining differs from bulk micromachining in that a deposited film acts as the structural layer, rather than the substrate itself. The process flow for surface micromachining is outlined in Figure 2.6. A sacrificial material is deposited onto a substrate and patterned. Subsequently, a film is blanket deposited onto the device to define the structural layer. After this, the sacrificial material is removed, leaving behind the desired microstructure [26]. The process requires a compatible set of structural materials, sacrificial materials, and chemical etchants with high etch selectivity such that etching of the sacrificial material does not affect the structural material [16]. It is required that the sacrificial materials have good mechanical properties to avoid problems in the fabrication process [16]. Finally, the structural material must offer the appropriate chemical and mechanical properties for the desired application.
Fig. 2.6 Surface micromachining process.
Materials and Microfabrication Processes
47
The polysilicon and silicon dioxide material combination is commonly used for the fabrication of microfluidic devices. The oxide is deposited onto the substrate and acts as the sacrificial material, with the deposited polysilicon acting as the structural material [26]. The deposition of polycrystalline silicon is typically achieved by gas-phase decomposition of silane using low-pressure chemical vapor deposition at temperatures in the range of 585 - 625oC. The wafer is washed using an HF solution in which the oxide is dissolved without affecting the polysilicon structure [26]. Conversely, silicon nitride can also be used as the structural layer with phosphosilicate glass (PSG) and silicon dioxide acting as the sacrificial material [27]. A combination of surface micromachining, bulk micromachining, and doping technologies has been used for the development of silicon processed microneedles by Lin et al. [28]. The process begins with a lightly doped ptype [100] oriented wafer which is selectively doped to make a 12 µmdeep heavily boron doped p-type region using a thermally grown oxide that is patterned to act as a masking layer. When the doping is completed, the oxide mask is stripped and a 400-nm-thick layer of SiO2 is thermally grown and a 600nm layer of nitride is deposited for passivation of the surface. A 600nm thick phosphorus-doped polycrystalline silicon layer is then deposited, patterned and etched, leaving behind the polysilicon resistors. After this, the polysilicon layer at the back side of the wafer is etched away and a 150nm protective layer of nitride is deposited to cover the polysilicon resistors during EDP etching. These steps prepare the substrate for bulk micromachining at the end of the process flow. Following these steps, the surface micromachining fabrication of the microchannel is performed. First, a 5µm thick layer of PSG is deposited, followed by a 3µm layer of Low temperature oxide (LTO), which improves adhesion to the photoresist and shields the interface from HF. Using a photoresist, which is subsequently stripped, the microchannel is patterned etched in buffered HF, leaving the cross section shown in Figure 2.7 (b). To create etch openings for the HF to remove the sacrificial PSG and LTO inside the microchannel, a 1µm layer of LTO is deposited to allow patterning of the openings. A 1µm layer of nitride is then deposited and patterned to create the etch holes. These holes are etched in a plasma etcher, providing the cross section as shown in Figure 2.7 (c). The etch channels are sealed by the deposition of a 1.5µm thick layer nitride after the microchannel has been fully cleared. In order to enable the separation of the needle, etch windows are opened using a plasma etcher which stops at the 400-nm-thick oxide layer. This is shown in Figure 2.7 (d).
48
Noori, Upadhyaya, and Selvaganapathy
(a)
(b)
(c)
(d)
(e)
Fig. 2.7 Process flow for fabrication of silicon processed microneedle (Reproduced with permission from [28] © [1999] IEEE).
Subsequently, the back side of the wafer is now patterned, which opens up the etching areas to free the microneedles. A timed EDP etch is used to reduce the silicon wafer thickness to 120µm, after which the wafer is immersed in a buffered HF solution which etches the opened SiO2 areas as shown in Figure 2.7 (e). Again the wafer is etched in EDP to reduce the 120µm thickness down to 50µm at the shank end of the microneedle. This fully releases the needle and completes the fabrication process, with the final results show in Figure 2.8.
Materials and Microfabrication Processes
49
Fig. 2.8 Silicon processed microneedle (Reproduced with permission from [28] © [1999] IEEE).
2.3 Glass Based Materials Glass is the most suited material for biological applications, primarily because of its favorable material properties (chemically inert, excellent insulator, optically transparent, and low auto-fluorescence), its close association with the life sciences, and the availability of virtually an infinite variety of glasses. Table 2.2 provides some of the important material properties of glasses mainly used in microfluidics. However, microfabrication techniques for glass have not advanced as much as those for silicon and it is difficult to produce high aspect ratio anisotropic structures in glass due to its amorphous nature. Early examples of microchannels constructed in glass were used as electron multipliers in X-ray applications [29] and as micropipettes for patch clamp applications [30]. Photosensitive glass has also been used in these applications to form high aspect ratio microchannels through UV exposure through a mask [31]. These methods involve heating glass capillaries and applying tensile force to draw a fine filament in micrometer dimensions and hence are not amenable to the parallel fabrication that is typical of photolithography. Subsequently, in the 1980s, unstructured thin glass was used as an actuating membrane in several silicon micropumps [32]. Actual structuring of glass using lithography for creating a network of microchannels, which is typical of microfluidics these days, began with its application in electrophoresis chips in the early 1990s [33]. Since glass capillaries were used widely in capillary electrophoresis and all the protocols had been standardized on them, early microfluidic systems were made of glass to replicate the surface characteristics and hence performance of macrosystems. Subsequent to this, development of a number of microfluidic devices
50
Noori, Upadhyaya, and Selvaganapathy
for diverse applications such as flow cytometry [34], micellar electrokinetic capillary chromatography [35], protein pre-concentration [36], electrospray coupling for mass spectrometry [37], and integrated systems for amplification, separation and detection of genetic materials [38] were made using glass as the structural material. A number of companies such as Micralyne Inc., Micronit BV, Innovative Micro Technology (IMT), LioniX BV, offer foundry services or produce limited volume of several microfluidic devices in glass. Table 2.2 Material properties of commonly used glass based material in microfluidics Properties
Soda lime
Electrical Dielectric strength 10 (MV/m)
Quartz
Pyrex Foturan (Corning 7740)
8
14
1.2
Resistivity (ohm-cm)
4x106
7.5x1015
8x1010
8.1x1012
Dielectric Constant
7.3
3.75
4.6
6.5
20
48
20.7
60
64
78
0.20
0.22
2.23
2.37
Mechanical Tensile strength (MPA)
Young’s modulus 70 (GPa) Poisson’s ratio
0.25
86.7 – 107.2 (direction dependent) 0.17
Density (g/cm3)
2.5
2.2
Materials and Microfabrication Processes
51
Table 2.2 (cont.) Properties
Soda lime
Thermal Thermal Conduc- 1.12 tivity (W/m°C)
Quartz
Pyrex Foturan (Corning 7740)
1.4
1.1
Melting Tempera- 700°C (soften- 1750 °C ture ing) 9.2 x10-6 /C Coefficient of thermal expansion Specific heat 0.21 (J/g.K) Other Properties Index of refraction 1.52
1.35
0.54 x10-6 /C
820°C (soften- 465°C (transing) formation temp.) 3.3 x10-6 /C 8.6 x10-6/ C
0.72
0.75
0.88
1.47
1.47
1.515
2.3.1 Microfabrication in glass Microstructuring of glass is mainly accomplished by bulk micromachining using isotropic wet etching. This process produces an isotropic etch, the aspect ratio of which can be controlled by the amount of stirring present in the etching chamber. This creates the pattern of the microfluidic network on the glass substrate. Fluidic access holes at reservoirs are created by a number of methods including, laser micromachining, ultrasonic drilling, abrasive powder machining, and drilling. Subsequently, the processed wafer is bonded with another glass wafer through anodic or fusion bonding to form a closed microfluidic network. Functional materials such as electrophoresis gels, porous polymers for chromatography, environmental sensitive hydrogels, and phase change actuation materials are incorporated in the microfabricated channels individually in post-processing. 2.3.1.1 Wet isotropic etching
Glass is generally indestructible to chemical attack from a wide range of chemical agents. However, a few chemicals such as hydrofluoric acid, concentrated phosphoric acid (when hot, or when it contains fluorides), hot concentrated alkali solutions, and superheated water, are capable of etching glass. HF is the most powerful of this group, etching any type of silicate glass. Other agents have significantly lower etch rates and require aggressive conditions and hence are generally not considered for controlled
52
Noori, Upadhyaya, and Selvaganapathy
etching. The etching due to HF is attributed to the presence of fluorine containing species F-, HF, HF2- and is depicted by the following chemical reaction: SiO2 + 6HF H2SiF6 + 2H2O
(2.1)
Of these radicals, the bifluoride (HF2-) is considered to have the greatest effect on the etch rate [39]. Etch rate has been shown to increase with HF concentration in the etchant mixture and vary with the inclusion of alkali oxides in its composition [40]. In contrast to glass, quartz has crystalline structure and hence exhibits anisotropic etching much like silicon with the same etching mechanism as described above for glass [41]. Z-cut (Z axis is the optical axis in quartz crystal and the unit lattice is symmetrical around this axis) wafers are used as the etch rate in that direction is greater that any other direction. The etched profiles vary greatly with process conditions, most notably the etchant concentration and temperature. Much like silicon, quartz has several crystalline planes making prediction of anisotropic etch profile difficult. However, several experimental [42] and theoretical [43, 44] studies on etch rates and profile development have been performed and can be used to design masks for obtaining appropriate shapes required for microfluidic structures. HF wet etching in conjunction with various other microfabrication techniques such as metallization, chemical mechanical polishing, and bonding has been used widely to fabricate microfluidic devices. For example, Lagally et al. [38] used wet etching of glass substrates to fabricate an on-chip PCR chamber. As shown in Figure 2.9 the glass substrate was coated with a thin layer of poly-silicon, which served as an etch mask, using a chemical vapor deposition process. The poly-silicon layer was patterned using a CF4 plasma dry etching with a photoresist layer patterned using photolithography as the etch mask. Another popular masking material is metallic multilayers such as Chromium/Gold (Cr/Au) that is sputtered or evaporated on glass substrates and patterned using photolithography [33]. Gold, due to its inert nature, serves as a suitable mask in preventing HF attack on the glass underneath. Chromium serves as an adhesion layer in preventing delamination of Au from the surface. In some instances the photoresist that is used to pattern the Cr/Au layer is left on during etching as additional protective layer to prevent etching through pin holes in the metal layer. Alternatively, thick negative photoresists such as those based on novolak resin or SU-8 have also been used as etch masks through a single photo-
Materials and Microfabrication Processes
53
lithography step, thus reducing the complexity and cost associated with fabrication [45].
Fig. 2.9 Process flow for fabrication of microchannels in glass.
This patterned substrate was subsequently etched in a 1:1:2 HF:HCl:H2O mixture at an etch rate of 6 µm/min to obtain a final etch depth of 42 µm and a channel width of 100 µm. Soda lime glass is a multi-component mixture consisting of small amounts of Na2O, CaO, MgO and Al2O3 apart from SiO2. Etching of this material using concentrated HF or buffered HF solutions causes precipitate formation which mask subsequent etching of the underlying regions leading to increased surface roughness. Addition of HCl was found to dissolve these precipitate and produce a smoother finish [46]. Subsequently, the photoresist layer was removed and the remaining poly-silicon etched using a CF4 plasma dry etch. Access holes (with diameters of 2.5 mm and .75 mm) to the microchannels were drilled using Diamond tipped drill bits to a depth of 1 mm. The etched and drilled substrate was thermally bonded, at 560 °C for 3 h, to a 210 µm-thick flat glass wafer of identical radius in a programmable vacuum furnace. Tygon tubing of 1/8 inch outer diameter was used to for fluidic connections.
54
Noori, Upadhyaya, and Selvaganapathy
2.3.1.2 Plasma dry etching of glass
The working principle of RIE and DRIE for glass is similar to that used for Si. However, the process is complicated due to the multi-component nature of glass. Various gas phase etchants such as CHF, CHF3 and CF4 that produce fluorine free radicals in the plasma and are directionally accelerated to the substrate are used for RIE etching [47]. The resulting fluorine radicals react with silica and produce volatile reaction products in the form as shown in the equation: SiO2 + CF4 CO2 + SiF4
(2.2)
Etch rates for soda lime glass of about 60 Å/min at 300 W and 60 mTorr pressure of CF4/Ar were obtained [47]. The etch rate was shown to vary with the chemical composition of the etchant mixture and the nature of the glass. This relatively slower rate of etching is said to be due to deposition of non-volatile compounds formed during reaction with glass components [48]. Higher etch rates of 0.2 - 1 µm/min for quartz and 500 nm/min for Pyrex have been obtained using inductively coupled plasma with SF6 as the gaseous etchant at pressures of 2-10 mTorr and a high bias of 340 V [49, 48]. 2.3.1.3 Photo-definition of glass
Certain types of glasses can be photostructured using UV light such that they crystallize in the exposed regions and the etch rate is increased significantly. In this way anisotropy induced by UV [50] or laser exposure can be used to fabricate high aspect ratio structures. Such glasses, also known as photoetchable glass, typically contain traces of rare earth metal oxides in them [50]. An example is a lithium aluminum silicate glass, also known commercially as FOTURAN (Schott Glassworks) that contains CeO2, Ag2O and Sb2O3 as rare earth impurities. When the glass is melted to form the wafers some Ce3+ ions are formed that are stabilized by Sb2O3, or other reducing agents: 2 Ce4+ + Sb3+ ↔ 2 Ce3+ + Sb5+
(2.3)
2
On UV exposure at 310 nm (dose of 2 J/cm for 1 mm thick plate) the Ce3+ ions oxidizes reducing to Ce4+ and releasing an electron: Ce3+ + hν (312 nm) Ce4+ + e-
(2.4)
This electron reduces Ag ions to Ag metal atoms: +
Ag+ + e Ag
(2.5)
Materials and Microfabrication Processes
55
Subsequent heat treatment at 500 °C causes agglomeration of silver atoms and formation of larger nuclei. The temperature is then increased to 600 °C when the surrounding glass region crystallizes around the nuclei forming lithium-metasilicate Li2SiO3 having a different chemical and physical properties than the unexposed glass region. The photodefined substrate is etched using 10% HF solution. An etch rate of ~10 µm/min; a selectivity of ~20; and a minimum line width of 25 µm with aspect ratio of 20 were demonstrated with a roughness less than 1 µm [50]. 2.3.1.4 Fluidic interconnect fabrication
A number of methods are used for creation of fluidic feedthroughs to connect external macrodevices such as pumps and valves with the microfluidic network. They include ultrasonic machining, laser micromachining, abrasive powder machining, micro electrochemical discharge machining, and conventional drilling [51]. The simplest method of machining glass is using single crystal diamond or poly crystalline diamond tools in ductile mode machining at low depth of cuts to precisely define microchannel shapes or for through hole machining. Brittle materials such as glass are difficult to machine mechanically due to fracture generation, tool breakage, and wear. However, under certain controlled conditions such as low depth of cut, it is possible to machine brittle materials such as ceramics using single or multipoint diamond tools so that material is removed by plastic flow, leaving a crack-free surface [52]. The transition from a brittle to a ductile mode can be thought of as a balance between the strain energy and the surface energy to cut a chip. In brittle materials the density of defects is small and therefore when the thickness of the chip cut is small, the stress field generated is also small that it does not encounter any defect and thus does not generate fracture. Poly crystalline diamond tools with randomly distributed sharp protrusions of diamond with dimensions around 1 µm have been used to machine grooves, pockets, and through holes in soda lime glass. The tools are made of a shank of tungsten carbide which is machined using a wire EDM method with surface roughness of 0.3 nm. Abrasive powder machining for glass in the microscale was initially developed at Philips research laboratories for through hole interconnects in glass sheets for flat panel displays [53]. The mechanism is based on material removal due to generation of micro-cracks by sharp indenting particles [54]. The process involves three basic steps. The first is masking using electroplated metals [55] and abrasion resistant polymers such as urethanes
56
Noori, Upadhyaya, and Selvaganapathy
[56] to protect the regions that are not targeted for etching. The second is erosion of the exposed parts of the wafer by high velocity jet of abrasive particles. In this step, the size of the abrasive particle has to be chosen small enough to enter the mask openings for effective erosion of the substrate. The size of the abrasive particles, along with the other parameters such as kinetic energy of the particles, determines the depth of cut [57]. The abrasive particle dispensing system provides a constant flow rate of particles with constant kinetic energy uniformly over large surface area. The third step is removal of the masking layer and cleaning of the substrate. Microfluidic channels of 100µm width have also been machined in glass [58] with 30µm alumina particles and resolutions down to 10µm have been achieved with smaller abrasive particles [59]. Electrochemical discharge machining (ECDM), otherwise also described as electrochemical arc machining, spark-assisted chemical engraving, and electrochemical spark machining, is a low-cost method to obtain deep anisotropic features with surface roughness similar to that obtained by HF etching. In this process, the substrate to be machined along with the two electrodes (cathode and anode) are immersed in an electrolyte solution (typically, sodium hydroxide, or potassium hydroxide). The cathode which is connected to the tool is placed close to the substrate. The anode, typically a platinum wire, is placed far away and has a larger surface area than the tool electrode. When the applied voltage is higher than the critical voltage [60], electrochemical reaction results in generation of H2 bubbles at the cathode, temporarily creating a gas film, separating the cathode from the electrolyte. This results in the buildup of voltage across this film and electric discharge across it. The etching is due to the thermal effects of the discharge combined with the chemical corrosion due to the electrolyte accelerated at that spot. Use of a rectangular pulse and tool rotation was found to improve the machining quality and obtain straight sidewalls. Rectangular pulse voltages of 40 V with tool travel rate of 1000µm/min and rotation of 1500rpm has been used to machine grooves of 367µm depth and 200µm width in pyrex glass [61].
2.4 Wafer Bonding For the construction of complex devices it is sometimes necessary to assemble individual layers for the final device. Planarization and whole wafer bonding methods can be used to assemble individual layers to obtain three-dimensional structures that are thicker than a single wafer and this
Materials and Microfabrication Processes
57
can be done simultaneously for several devices. In microfluidic devices, bonding is used primarily to provide a top cover to open microfluidic channels and networks formed using either silicon or glass using bulk micromachined methods. The other substrate to be bonded could be of silicon, glass, or even other metals and can be instrumented with electrodes, heaters, sensors, detectors etc. to increase the functionality of the device. Several methods have been used for this purpose, including fusion bonding, anodic bonding, and adhesive bonding methods. Bond strength, maximum pressure to failure, low chemical reactivity of interface materials, and high temperature stability are the key factors in the choice of approach. A good review of wafer bonding techniques and metrology is described elsewhere [62] 2.4.1 Fusion bonding Several processes have been developed, with fusion bonding being the most common [16]. In wafer fusion bonding, the silicon wafers are bonded at room temperature where they adhere to each other by hydrogen bridge bonds that develop between the OH terminated surfaces across the interface. The energy of this bond is low compared to that of covalent bonding. The wafers are then annealed at temperatures between 700°C to 1100°C that causes the Si-O-Si bonds to react [16]. Initially as the temperature is ramped up, water molecules at the interface rearrange and produce stable hydrogen bonding at temperatures below 110°C. Further increase in temperature forms stable siloxane bond at the interface at temperatures between 150 and 800°C. Finally, viscous flow of interface oxide occurs at temperatures higher than 800°C. Tensile strengths as large as 9MPa have been obtained using this method. A major problem with wafer fusion bonding is the presence of non-contacting areas on the silicon wafers, which can be caused by particles, residues, and surface defects. Therefore, fusion bonding requires that surfaces are perfectly smooth as small particles can cause a large void [16]. Another major issue is incompatibility of electronic circuits and semiconductor detector that have been fabricated on the wafers to high temperatures that occur in this process. 2.4.2 Anodic bonding Anodic bonding is one of the most widely used methods for bonding of glass microfabricated substrates to silicon, metal, or other glass wafers. The term was coined by Wallis and Pommerantz in 1969, when they observed that by applying an electric field a bond between metal foil and
58
Noori, Upadhyaya, and Selvaganapathy
glass could be achieved at relatively low temperature [63] compared to fusion bonding. This is ideal for bonding substrates containing electronic circuits, sensors and actuators or substrates with large difference in thermal coefficient of expansion. Typical operating conditions are 200-1000V and moderate temperatures [300-450°C]. This method has extensively been used in microfabrication of microfluidic devices such as micropumps and microsyringes [64, 65], as well as in other MEMS applications like inertial sensors [66] and pressure sensors [67]. In microfluidics, the microchannel network is sometimes structured in glass and bonded to a silicon wafer using anodic bonding. At elevated temperatures of 300-4500C, the mobility of ions in the insulator is increased [68] resulting in transport of mobile Na+ ions away from the interface. This causes formation of a depletion zone with excess O- ions that are chemically bonded at this elevated temperature to the Si ions in the adjoining wafer [69, 70]. The electrostatic attraction between the two wafers at high electric fields pulls together the interface [71] and hence the process is more tolerant of small surface roughness compared to fusion bonding. The same mechanism explains bonding at the glass-glass and glass-metal interfaces [72]. Glass-glass bonding involves additional step of creating an intermediate layer using dissolved wafer process where a silicon wafer is bonded first to a glass wafer and then etched back to retain a SiO2 layer. This modified wafer is bonded to the structured glass substrate using the same process as described for glass-silicon bonding. Alternatively a layer of silicon nitride, oxide, or carbide can also be used instead of the dissolved wafer process for creation of the intermediate layer. 2.4.3 Adhesive bonding This technique of using an intermediate adhesive layer to bond wafers of similar or dissimilar materials has been in existence for a long time. Typical intermediate layers are polymers such as polyimides, epoxies, thermoplast adhesives, elastomers such as PDMS and photoresists. They are spin coated to one of the substrate to produce a thin uniform layer prior to bonding. Then depending upon the adhesive material relatively low curing temperature (100°C) and moderate pressure is applied to bond them [73]. The main advantages of this technique over anodic and fusion bonding are: low cost, low bonding temperature and greater tolerance to surface roughness and topography. This process does not produce hermetically sealed structures, the life span of polymers is low, and polymers are not suitable for high temperature processes [74]. Formation of voids due to entrapped
Materials and Microfabrication Processes
59
air and solvent evaporation can also be an issue [75]. The main factor in establishing good bonding strength is absence or reduced release of reaction products in the vapor form during the curing process. In this context PDMS prepolymer has been used as the intermediate layer and transferred on to the bonding surface using imprint lithography [76].
2.5 Polymers Polymers are macromolecules with a high molecular mass and a molecular structure consisting of repeated monomeric units. They can be classified into two categories, linear and cross linked polymers, depending on the position of the reactive groups in the monomer and the cross-linker. The polymerization process is statistically dependent and hence a range of polymer chain lengths with cross-linking is created forming a large 3-D network. Hence polymers usually do not have a defined melting temperature but soften over a range called the melt interval. A wide variety of polymers have been created from a number of starting monomers having different physical and chemical properties [77]. Polymers, depending on their response to thermal treatment, have been classified as thermoplastic, thermosetting, and elastomeric polymers. Thermoplastics are mostly linear or branched polymers that are melted upon the application of heat and re-solidify when cooled. They do not have cross linking and hence their thermal behavior is reversible. At a temperature above the glass transition temperature these materials become plastic and can be molded into specific shapes, which they retain at lower temperatures. Thermosetting polymers are heavily cross linked and hence the molecular movement for elasticity is not possible. They are rigid and brittle and do not soften significantly with increased temperature. Elastomers are weakly cross linked polymers. They can be easily stretched and will revert back to their original shape upon release. Since they are cross linked they do not melt before reaching their decomposition temperature. 2.5.1 Microfabrication Although silicon and glass have been the materials that have been primarily used for microfluidic devices, they have a number of disadvantages. Silicon based devices have high substrate electrical and thermal conductivities and therefore modification in design and additional processing steps are typically required for their use in microfluidic systems. (For ex-
60
Noori, Upadhyaya, and Selvaganapathy
ample, insulation layers are needed to prevent a short circuit through silicon substrate in electrokinetic devices. Similarly, isolation islands have to be incorporated in the design for microreaction chambers to prevent thermal cross talk. Glass provides excellent isolation properties but is very difficult to machine into high aspect ratio complex 3-D structures. Polymers in contrast are moldable to complex shapes and dimensions and have comparable isolation properties to glass. Polymers also have a number of attractive properties such as low cost, chemical inertness, low electrical and thermal conductivity, suitability for surface modification and compatibility with biological materials that make them ideally suited for use in biological microfluidic applications [78]. Some of the common thermoplastic polymers include polymethylmethacylate, poly-carbonate, poly-vinylidenefluoride, poly-sulfone, poly-styrene, poly-vinylchloride, poly-propylene, poly-etheretherketone, polyoxymethylene, and poly-amide. These polymers cannot be microstructured using the traditional microfabrication techniques. However other techniques such as injection molding, hot embossing, and casting can be used for microfabrication of polymers by exploiting their compliant form at high temperatures. In these methods, a substrate made of the polymer is fashioned as a negative replica of the master, which retains its structural integrity at these temperatures. Once a master mold has been made, typically using traditional microfabrication techniques, several thousand polymer replicas can be molded with low cost. The replication process is insensitive to the complexity and fineness of the design, and can be performed outside the clean room while still obtaining device resolution similar to photolithography. This aspect of polymer micromachining reduces the cost of the devices significantly – making the technology ideal for disposable biomedical devices [79]. Surface micromachining techniques have also been recently adapted to suit low temperature processing for polymeric materials. These allow construction of devices with multiple materials and complex functionality. Outlined below is a brief description of the different methods widely used to structure 3-D objects in the microscale with polymer substrates. Several companies such as Micronics, Steag microParts, MicroTec, Amic AB, Micralyne, and DALSA provide foundry services for fabrication of microfluidic chips in polymers. 2.5.1.1 Injection molding
Injection molding has been one of the first techniques adopted for micro manufacturing of polymeric structures for microfluidic devices. It is a well established technique and the manufacturing infrastructure, technology and
Materials and Microfabrication Processes
61
process know-how already established have been rapidly adapted for microscale molding [80]. Several groups have demonstrated the feasibility of replication in thermoplastic polymers of microstructures [81, 82, 83, 84]. Similarly, a number of investigations have also been performed on fabrication of micromold inserts with microscale features. These include using microfabrication techniques such as silicon micromachining [9], SU-8 photolithography, [85] and other machining techniques such as micro EDM and LIGA process [86]. In injection molding the mold cavity with the mold inserts is evacuated and heated to a temperature higher than the glass transition temperature of the polymer. A horizontal injection unit heats the polymer and injects the viscous melt into the mold cavity. The entire cavity is then cooled below the glass transition temperature and subsequently the molding part is demolded. Figure 2.10a depicts the operation. The quality and precision of the mold insert determines the quality of surface features. Injection molding is the most suited for industrial production due to short cycle times (~ mins). 2.5.1.2 Hot Embossing
Hot embossing refers to a stamping process where microstructural features from a hot mold insert (master) is transferred on to a thermoplastic substrate. Figure 2.10b depicts the process. In this method, a thermoplastic film or substrate is inserted into the embossing mold with mold inserts on either side. The mold is heated to a temperature just higher than the glass transition temperature of the thermoplastic material. The tool is evacuated and the heated mold inserts are pressed against the thermoplastic film with high force. The thermoplastic material close to the surface, due to the mechanical force and temperature is redistributed in such a way as to form a negative replica of the feature on the mold insert. The setup is cooled and the mold insert is withdrawn from the plastic [79]. In injection molding, the material is injected and fills the cavity at high temperature. Subsequent cooling introduces significant stresses in the bulk of the material. Unlike injection molding, in hot embossing the thermoplastic material flows only over short distances, which results in low stress. Local heating and a low thermal capacity requirement also reduces the thermal time constant for embossing and delaminating and hence achieves a faster process time [79]. Since high aspect ratio structures (>10) can be replicated quite easily in thin films (1-100 µm), this technique has been adapted to replicate patterns on photosensitive films using a process called nano-imprint lithography.
62
Noori, Upadhyaya, and Selvaganapathy
Fig. 2.10 Process flow for various fabrication processes available for microstructuring polymers.
Materials and Microfabrication Processes
63
2.5.1.3 Casting
In this process a material is introduced into a mold in its liquid state, allowed to solidify in the shape inside the mold, and then removed producing a replica of the microstructured master. Figure 2.10c illustrates the process. Molds for casting are also created similar to those for injection molding and hot embossing. A number of materials including hot, liquid metals and thermoplastics have been cast. A variation of this process is what is termed as reaction casting. In this process a two component mixture consisting of a long chain polymer (base) and short chain polymer with initator (crosslinker) is used. The mixture in the uncrosslinked state is fluid and wet the surface attains the shape of mold insert when cast. Subsequent exposure to heat or UV light input coldsets the mixture through cross linking - increasing its viscosity by several orders of magnitude. This process is typically used for non-thermoplastic materials such as thermosetting plastics (bakelite), epoxies [87], and elastomers (PDMS) [88]. In this process the drawbacks are that complete mixing and distribution of the two components have to take place and the chemical reaction (cross linking) has to take place throughout the bulk for uniform strength. This can only achieved by comparatively long cycle times (~hours). However feature in the scale of a few nm can be easily replicated. 2.5.1.4 Stereolithography
Stereolithography is a rapid prototyping technology first developed in 1980’s [89] for prototyping in CAD/CAM applications. Figure 2.10d shows the schematic of the process. Here, a laser beam is focused on the free surface of a photosensitive liquid to induce polymerization of the local region of liquid and to transform it to a polymerized solid. The object to be formed is deconstructed into series of two-dimensional layers and reproduced by laser induced polymerization of a resin. Resolution depends on the laser wavelength, scanning apparatus used and the diffusion in the photosensitive liquid. Penetration depth of the laser light into the liquid determines the thickness of the layer. After laser exposure of one 2D crosssectional layer shape, fresh liquid is spread on top and the second layer is created similarly. The resolution of this technique is ~150-200 µm in three dimensions. This resolution has been improved with the use of a glass holder (through which the laser light transmits) attached to an X-Y-Z micro-positioner to accurately control the position and thickness of the resin with which the light interacts [90, 91] in the resin volume. Another method, free surface polymerization, was then developed wherein the part formed is attached to the X-Y-Z positioner and moved close to the free
64
Noori, Upadhyaya, and Selvaganapathy
surface of the resin/ air interface [92] achieveing micrometer resolutions. The throughput has been increased by using optical fibers to split the laser light and to manufacture several part simultaneously [93]. Liquid Crystal Display (LCD) devices [94] and Digital Micromirror Devices (DMD) [95] have been used to generate patterns and eliminate X-Y movement and provide a lithography type process which decreases the time to generate layers and increases throughput. An excellent review of stereolithography and its applications has been published [96]. 2.5.2 Polymer materials: Some of the most popular and widely used polymeric materials in microfluidics include Parylene (paraxylylene), PDMS (polydimethylsiloxane), SU-8, hydro-gels, acrylate based porous polymer monoliths, biodegradable polymers, polyesters, polyimide, and Paraffin. Some of these polymers such as parylene, PDMS, SU-8, biodegradable polymers, and polyimide are primarily used for the construction of microfluidic channels. Others including paraffin, hydrogels, and porous monoliths are used as functional material in the construction of valves, pumps and reactor beds. This section deals with the detailed description of the properties of these polymers and their fabrication processes. 2.5.2.1 Parylene
Parylene is the industry name for a class of polymers called poly paraxylene. It is available commercially in three dimeric stable forms, which are variations of a basic polymer backbone of xylylene with replacement of 1-4 atoms in the benzene ring with chlorine as shown in Figure 2.11. It has been used for a wide variety of applications such as encapsulation for microelectronic circuits [97, 98], as interlayer dielectrics [99], and for microchip packaging [100]. Similarly, a number of applications of parylene in microfluidics include microchannels [101, 102], microvalves [103, 104, 105], membrane filters [106], and other micromachined devices. Parylenes have to be polymerized in-situ and cannot be formed by extrusion or molding techniques because of their high molecular weight. Parylene polymers are currently deposited by vapor deposition and in-situ polymerized at room temperature and in vacuum. The dimer form of parylene, which is solid at room temperature, is sublimated at 140o-160oC and passes as a vapor into the pyrolysis chamber where it then splits into two monomers at 680oC. The reactive monomer flows into a deposition
Materials and Microfabrication Processes
65
chamber where it deposits and begins to polymerize on surfaces below 100oC. The deposition pressure is ~ 100 mTorr and hence the deposition is conformal. This allows for the formation of uniform conformal thin films.
Fig. 2.11 Molecular structure of various commercially available parylene polymers.
Parylene thin films are extremely conformal even with high aspect ratio structures due to the vapor based deposition process. Low surface roughness, pinhole free coating, and excellent dielectric breakdown characteristics can be obtained for thickness greater than 0.5 µm. The films also are chemically pure due to lack of initiators and catalysts in the polymerization process. They are stress free due to room temperature polymerization process. The polymerized films are clear and possess low auto fluorescence allowing visualization of biological and chemical interactions through optical means. They are chemically and biologically inert and biocompatible. Hence they are resistant to damage by acids, bases, corrosive
66
Noori, Upadhyaya, and Selvaganapathy
body fluids, electrolytes, enzymes, and proteins. Parylenes have superior barrier properties compared to silicones and other polymeric materials. Their moisture vapor permeability is 1.7 x 10-16 kg-m/N.s., which is an order of magnitude lower than silicones. They resist room temperature chemical attack and are insoluble in all organic solvents up to 150° C. Parylene C can be dissolved in chloro-napthalene at 175° C, and parylene N is soluble at the solvent's boiling point (265° C). Substituted parylenes have been developed recently that can react with biomolecules to promote strong attachment [107]. A surface micromachining process for construction of Parylene microchannels has been demonstrated [108]. The process begins with the initial coating of the substrate with a thin conformal parylene layer. Adhesion to the substrate is usually obtained with plasma treatment and silanization by vapor deposition. Next, electrodes and other functional parts of the microfluidic device can be deposited and patterned. Thick (20 µm) photoresist is spun cast and photo-lithographically patterned to define the microchannel shape and dimensions. A second layer of parylene is then conformally deposited to form the top and sides of the channel. The adhesion of this layer is also assisted by a short oxygen plasma treatment. Using a thick photoresist mask, reservoir, and contact pad openings are etched using oxygen plasma. Dry etching in O2 and CF4 atmospheres at low pressures produces etch rates of 0.2 µm/min at 250 W and 50 mTorr [109]. The sacrificial photoresist in the microchannel is then dissolved in acetone for 36 hours without agitation and rinsed and dried. The release time could be shortened by constant agitation or heating the acetone. A number of devices such as DNA separation devices [108], PCR systems [102], cochlear implants [110], microvalves for drug delivery [104, 105], and microneedles have been fabricated using these techniques 2.5.2.2 SU-8
SU-8 was developed by IBM [111] for LIGA applications. A modified version of this material was used to produce high aspect ratio structures for MEMS applications using conventional lithography [112]. Although positive photoresists have been used widely as sacrificial layer for surface micromachined construction of microchannels [102], SU-8 was one of the first negative tone resists to be used as a structural layer for MEMS and microfluidic applications. It consists of EPON SU-8 resin (an epoxy) as its main constituent with γ-butyrolacton as the solvent and 10% wt of photoinitiator. The versatility of this resist was demonstrated by the fact
Materials and Microfabrication Processes
67
that thicknesses from 750 nm to 450 µm can be obtained [112]. It is commercially available from MicroChem Inc. Processing and lithography of SU-8 is similar to conventional negative tone photoresists. SU-8 is spun on silicon wafer and pre-baked at 95oC for a duration of 15 min for every 100 µm thickness. The resist layer is exposed in a UV aligner at a dose of 300-400 mJ /cm2 at 365 nm per 100 µm thick resist layer. The resin has very low UV absorption and hence uniform dose and crosslinking throughout the bulk of the resist can be achieved. It is developed using propyleneglycol monomethylether acetate solution and rinsed in iso-propyl alcohol. The resist can be stripped at this stage using hot 1-methyl-2-pyrrolidon solution. Further hard baking of the resist at higher temperatures 120-200oC for 30 min makes it resistant to metal etchants and other solvents typically used in microfabrication [112]. High aspect ratio structures (>18) with a range of thickness from 80-1200 µm, with conventional mask aligners have been demonstrated [113]. Higher aspect ratios (>60) have been achieved using x-ray sources for LIGA applications [114]. Another feature of SU-8 is that taller structures can be constructed by using multiple layers. A 100 µm layer can be spun, pre-baked and exposed and this can then be used as a substrate for the next layer. Multiple layers with alignment using different masks can produce complex 3-D structures. After the final layer has been exposed and baked, all the layers can be developed at the same time. Several methods for fabrication of microchannels using SU-8 have been reported as depicted in Figure 2.12. A simple method as depicted in Figure 2.12a is the use of a sacrificial positive tone resist to define the microchannel structure. SU-8 is cast over this sacrificial mold and the access holes are lithographically defined. Then the sacrificial resist is removed to create an open microchannel [115]. A variation of this method is depicted in Figure 2.12b. Here, a layer of SU-8 is spun, pre-baked and exposed into patterns of microchannels. Then a thin metal layer is deposited on top to protect unexposed regions from exposure in the bottom SU-8 layer. A second SU-8 layer is then spun, pre-baked and exposed to a pattern to create access holes for the microchannels. Subsequently the second SU-8 layer is developed, metal layer etched and the unexposed resist in the first SU-8 layer also developed to produce the channel structure. [115]. A third method as shown in Figure 2.12c, depends on the energy of proton beam to define the penetration and hence exposure depth of the resist to produce three dimensional features. Initially a thick layer of SU-8 is spun cast. A
68
Noori, Upadhyaya, and Selvaganapathy
long exposure to 2MeV proton beam through a mask is done to produce a pattern of the microchannels throughout the entire depth of the resist layer that defines the walls of the microchannel. Subsequently the wafer is flood exposed to 0.6 MeV beamwith a penetration depth of 10 µm that polymerizes the entire top layer forming the top surface of the microchannel. Subsequently the unexposed regions are developed to produce the embedded microchannel structure [116]. The fourth method is depicted in Figure 2.12d. Here, the SU-8 layer is spun, pre-baked, exposed, and developed to produce a pattern of the microchannel. A dry Riston photoresist film is laminated over the SU-8 layer. Access holes are then patterned in the Riston layer to provide fluidic connection to the SU-8 microchannel [117].
Fig. 2.12 Methods for fabrication of microfluidic channels in SU-8.
Materials and Microfabrication Processes
69
2.5.2.3 PolydiMethylsiloxane (PDMS)
PDMS is one of the most widely used polymers in microfluidics and BioMEMS. It belongs to a class of polymers called silicones containing a Si-O backbone. To this backbone, organic groups are frequently attached to the silicon atoms via a Si-C bond. PDMS has a repeating (CH3)2SiO unit. Depending upon the number of repeat units in the polymer chain and the degree of cross-linking, the material can be obtained in different rheological forms such as fluids, emulsions, lubricants, resins, elastomers, and rubbers [118]. The low rigidity of the backbone allows the methyl groups to be easily exposed, resulting in low intermolecular interactions of PDMS and also its low surface tension. PDMS in its elastomeric form is used generally for microfluidic and BioMEMS applications. The commercially available version is provided as a kit (Sylgard 184 from Dow Corning and RTV615 from GE) with the PDMS base solution and a curing agent to cross link siloxane base oligomers (in the base) containing vinyl terminated end groups (CH2=CH-) into elastomer by an organometallic crosslinking reaction. The base solution also contains a platinum based catalyst and silica filler. Typical crosslinkers are dimethyl methylhydrogen siloxane or tetrakis(dimethylsiloxyl silane) (TDS). The curing agent also contains an inhibitor (tetramethyl tetravinyl cyclotetrasiloxane). The platinum-based catalyst catalyzes the addition of the SiH bond across the vinyl groups, forming Si-CH2-CH2-Si linkages. The multiple reaction sites on both the monomer chain and crosslinking chain enable three-dimensional crosslinking. A casting process is used to form microfluidic channels and structures using PDMS elastomer as the structural material [119]. The master molds for this process with high aspect ratios are created using wet chemical etching of lithographically patterned silicon substrate. Alternatively, SU-8, and poly-urethanes have also been used to form the master molds [120]. A mold release layer of 3% (v/v) dimethyloctadecylchlorosilane in toluene is vapor deposited for 2 h to facilitate delamination of the PDMS replica. The PDMS base and its curing agent, mixed in 10:1 ratio, are cast over the mold. The PDMS pre-polymer mixture, due to its low surface energy, wets the entire surface down to the molecular level forming an accurate negative replica of the relief. Curing of the PDMS prepolymer for 4 h at 65 oC initiates crosslinking reactions, which solidify the material into an elastic monolith. Subsequently, the PDMS replica is peeled off the master. The master can be used to create several PDMS molds, thus replicating micro and nanoscale features without the use of expensive clean room processes.
70
Noori, Upadhyaya, and Selvaganapathy
Access holes to reservoirs ae punched through the bulk material using a sharp needle. The device is then placed on a thin slab of PDMS in order to form a closed channel system of four equivalent walls. A hermetic seal was obtained by mere contact without applying large external force [119]. The chips manufactured using this technique withstood pressures of 1 bar. Buffer solution and separation media were pipetted into the reservoirs, and the microchannels were filled by applying vacuum to one of the reservoirs. Capillary electrophoresis separation of φX-174/HaeIII DNA was accomplished in this device, to demonstrate the feasibility of microchannel construction and its use in BioMEMS [119] Although adhesion between PDMS layers was obtained in the previous process, the bond was reversible. A method to irreversibly bond PDMS to a substrate in order to construct devices to withstand higher pressure has also been developed [120]. In its natural state the PDMS surface is inert due to hydrophobic CH3-groups on the surface, resulting in low adhesion. Surface oxidation of the PDMS in plasma removes some of the methyl groups and exposes the PDMS backbone containing hydroxyl groups. Oxidized PDMS surfaces brought into conformal contact with each other form a tight, irreversible seal, improving adhesion of the substrates and pressure tolerance of microchannels. The Si-OH groups at the surface of the two PDMS slabs form a covalent Si-O-Si bond between them. A prolonged plasma treatment (>30 seconds) however, results in the oxidized layer becoming thinner and cracking. If the treated samples are left to age in ambient atmosphere the hydrophobicity recovers due to low molecular weight PDMS migrating to the surface [121]. Although plasma oxidation is most popular, other techniques such as corona discharge [122] and UV light exposure [123] have also been used to reduce hydrophobicity, particularly to reduce protein adsorption in microchannels. Microvalves and pumps have been incorporated into PDMS microfluidic networks by modifying the fabrication technique described above to produce multiple stacked layers with microscale features embedded in them using a technique known as “multilayer soft lithography” [124]. In this process, several elastomeric layers consisting of microchannel networks and PDMS membranes are fabricated separately, aligned, assembled and bonded. This technique exploits the two-component curing process to obtain bonded PDMS interfaces. The bottom layer has an excess of one of the components (PDMS base), whereas the upper layer has an excess of the other (crosslinker). When these two layers are brought in contact with each other and thermally cured, crosslinking reaction occurs between excess materials in these layers forming covalent linkages and a hermetic
Materials and Microfabrication Processes
71
seal. The strength of the interface equals the strength of the bulk elastomer [124]. Additional layers are added by simply repeating the process with the polarity of material excess reversed. Valves and pumps using the pneumatic control have been demonstrated by constructing a two microfluidic layers sandwiching a membrane layer with the top layer being the pneumatic control network and the bottom layer being the fluidic network [124]. 2.5.2.4 Polyimide
Polyimide has also been used as a structural material for the construction of microfluidic devices such as microvalves and pumps [125, 126] due to its excellent chemical and thermal stability, low water uptake, and good biocompatibility. Commercially available polyimide resin consists of aromatic heterocyclic chains of alternating carbonyl groups, which act as electron acceptors and nitrogen groups which act as electron donors. This charge transfer complex stabilizes the chain and is responsible for its chemical inertness, structural rigidity, and high strength. Polyimides are made photosensitive by attaching a photosensitive group that is an inhibitor of polymerization. The photosensitive group detaches from the backbone upon UV excitation, initiating polymerization in that region [127]. A photosensitive commercially available polyimide (PI – 2732, Dupont) has been used as structural material for microchannel construction [128]. In this process, a thin polyimide layer is spun, pre-baked and patterned on a substrate using optical lithography with the desired geometry and vent channels, but not cured. Next, a thin layer of solvent with dissolved precursor is used to coat a soft-baked layer polyimide, on a second substrate. The two halves are placed in contact and cured. Channels with dimensions 50 to 1000 µm wide, 3 to 30 µm deep have been fabricated for chemical analysis and heat transfer devices [128]. The structures developed in this process have significant internal stress due to solvent evaporation during the final curing. To resolve this problem, a modified lamination method has been developed to create microchannels with embedded electrodes [126]. This lamination process uses a combination of photosensitive polyimide (PI – 2732) that is microstructured using optical lithography and non-photosensitive polyimide (Mylar film), which is microstructured using dry etching with oxygen plasma and a silicon dioxide mask. The PI-2732 polyimide is spun cast to obtain a 5 – 20 µm layer, exposed and cured at 350oC for 1 hour. This layer forms the base of the microchannel. A layer of Ti/Pt (50 nm /
72
Noori, Upadhyaya, and Selvaganapathy
200 nm) is sputtered and patterned using a positive photoresist. This forms the electrodes inside the microchannels. A second layer of PI – 2732 polyimide (5 – 20 µm), which forms the side walls, is spun, exposed and partially cured at 100-150oC for 1 hour. This layer not only insulates the electrodes but also provides the structure for microchannels. This layer is treated with n-methyl-2-pyrrolidone (NMP) (swelling agent) that causes the partially cured second layer to swell, enabling higher interdiffusion with other polyimide layers during lamination. A thin Mylar foil is spin coated with photosensitive polyimide, partially cured, flipped over, and bonded to the substrate by lamination. This layer forms the top of the microchannel. The Mylar layer is coated with oxide, access holes lithographically defined and dry etched in oxygen plasma [126]. No interface was observed between the open channel pre-structures and the laminated top layer, which indicates good channel sealing as shown in Figure 2.13. The channels thus fabricated were able to withstand pressures as high as 19 bar [126]. The curing results in shrinkage of the polyimide (30–40%), which degrades the sharpness of the sidewalls.
Fig. 2.13 Cross section of microchannels created in polyimide. Reproduced with permission from [126] © 2001 Royal Society of Chemistry.
2.5.2.5 Hydrogels
A gel is a loose network of cross-linked polymers that retains much of its solvent. Gels in which water is the solvent are called hydrogels. Any water-soluble polymer can be prepared as a hydrogel by carrying out the polymerization in the presence of a cross-linking agent in an aqueous solution of the polymer [129]. Hydrogels have been recently introduced in microfluidic devices to provide intelligent transduction of signal in one domain (biological, chemical, mechanical or fluidic) to another. Hydrogels can be engineered to be sensitive to a variety of environmental factors such as pH [130], temperature [131], electric field [132], glucose concentration [133], and antigens [134] and convert them to mechanical action such as expansion or contraction.
Materials and Microfabrication Processes
73
Diffusion of the stimuli into the bulk of the gel is the key factor in determining the response time [135]. Therefore, faster response can be achieved by reduction in size, making hydrogels attractive materials for microscale actuation. Since these gels have high solvent content, their properties depend upon the balance of polymer-polymer vs. polymer-water interactions. Structural change, size variation, and mechanical actuation can be accomplished by slight changes in the polymer-solvent interaction properties initiated by external stimulus of which temperature is the most common [136]. A linear polymer that displays cloud point behavior when heated can be crosslinked to give a temperature-sensitive gel network. For example, poly (Nisopropylacrylamide (polyNIPAM) or related copolymers, when heated undergo a rapid and reversible phase transition from extended hydrated chains to collapsed hydrophobic coils and shrinks by expelling water above a temperature termed as Lower Critical Solution Temperature (LCST) of 32 oC. These gels can be formed by photoinitiation, particularly with the use of UV light [137]. This feature makes them suitable for microfabrication through photolithography. A solution of Nisopropylacrylamide (NIPAM), allyl methacrylate and azobisisobutyronitrile (photo-initiator) in tetrahydrofuran (THF) solution was spun cast to 1 µm thickness and photo-lithographically defined using optical photolithography [138]. The solution can also be cast in a sacrificial mold on the substrate to obtain higher thicknesses. Photolithography and development in acetone produces microstructures with 50 µm resolution. The transduction property of this polymer has been used to fabricate an inline microvalve [138]. A 2.5-fold increase in polymer volume in water is observed when the water temperature changes from 25°C to 10°C as seen in Figure 2.14. Response time was reduced to 7 s primarily due to microfabrication of the hydrogel [138].
74
Noori, Upadhyaya, and Selvaganapathy
Fig. 2.14 Microfabricated hydrogel actuated using temperature.
pH sensitive hydrogels have also been used to regulate flow in microfluidic channels [139]. The photopolymerizable liquid consists of acrylic acid and 2-hydroxyethyl methacrylate (in a 1:4 molar ratio), ethylene glycol dimethacrylate (1 wt%), and a photoinitiator (3 wt%). In this method, PDMS microchannels were filled with the prepolymer and used to provide spatial resolution. This mixture was irradiated with UV light through a photomask to define the regions in the microchannel for polymerization [139]. Polymerization times varied depending on light intensity, photoinitiator, and monomer mixture, and was less than 20 seconds when Irgacure 651 was used as the photoinitiator. When the polymerization was finished, the channel was flushed with water to remove the unpolymerized liquid. In a different implementation, hydrogels responsive to glucose concentration were fabricated using a casting based approach [140] for use in microvalves. The expansion of the gel in response to glucose concentration regulates the flow in the microchannel situated below. This gel, also displays sensitivity to pH and demonstrates a 100% volumetric expansion to pH variation from 7 to 11 and a 30% volumetric expansion for glucose variation from 0 to 100 mM at physiological pH of 7.4. As noted earlier, the response time of the gel is dependent on its size and valve opening times as low as 7 min have been observed with 30 µm hydrogel layers [140]. 2.5.2.6 Macroporous polymers
Macroporous polymers are monoliths whose internal structure consists of an interconnected array of polymer microglobules separated by pores, and their structural rigidity is secured through extensive crosslinking [141].
Materials and Microfabrication Processes
75
These polymers have a fixed pore size, narrow pore size distribution, and due to extensive cross links are structurally stable with their porous structure persisting even in the dry state. These materials were initially produced in bead forms for applications as chromatographic and ion exchange column packings [142]. Phase separation of inert diluents (porogens) during polymerization was used as the mechanism for precise pore size control and its narrow distribution [143, 144]. They have also been demonstrated to be effective as monolith support for catalysis reactions [145]. The fabrication of these monoliths in molded fashion in various tube sizes and tube materials have been demonstrated [146, 147, 148]. These polymers have been used in high throughput bioreactors with immobilization of various catalytic enzymes on the pores [149], and are now finding applications in drug delivery, membranes, and encapsulation of bioreactors [129]. The surface properties of these pores formed can be changed from hydrophobic to hydrophilic by addition of grafted side chains [150]. The porous nature gives them high surface to volume ratio, which is desired in case of catalytic reaction, surface adsorption, chromatography, drug delivery, and electroosmosis. Microfabrication of these monoliths, has been demonstrated by using conventional lithographic techniques [151]. The main issues encountered were the low viscosity of the monomeric mixture and the effect of oxygen on the polymerization reaction. Low viscosity, restricts the ability of using spin coating to obtain defined uniform thickness of films. Very thin layers (~50nm) and non uniform distribution of polymer thickness were achieved by spinning. Secondly, oxygen is an inhibitor in the polymerization process. Thin layers of the monomeric solution exposed to atmosphere absorb oxygen, which inhibits polymerization. These issues were resolved using a microstructured glass mold which was etched to a depth of 20 µm and using it to enclose the monomeric mixture [151]. The substrate on which the polymer monolith is to be structured is used as the capping layer. Adhesion to the substrate was enhanced by depositing a monolayer of gammamethacryloxytropyl trimethoxy silane to it while it is reduced with glass mold by coating it with parylene [151]. Since the glass mold is optically transparent the monolith could be patterned using direct lithography. Light initiates radical generation of AIBN and polymerization of the monomers. Upon polymerization the polymer phase separates from porogen phase and polymerizes around the trapped porogens forming a porous structure. However, low viscosity causes diffusion of polymers and loss of resolution of the features created. Hence flood polymerization and using dry etching techniques for structuring creates high resolution structures as shown in Figure 2.15. These structures have been demonstrated to increase the
76
Noori, Upadhyaya, and Selvaganapathy
pumping capacity of micro electroosmotic pumps [151] and have potential applications in microscale bioreactors and filtration units.
Fig. 2.15 Macroporous polymers patterned on silicon wafers using conventional micromachining techniques.
2.5.2.7 Paraffin
Paraffins are saturated hydrocarbon mixtures, consisting of alkanes of varying chain lengths. The general chemical formula for straight chain paraffins is CnH2n+2 and have different melting temperatures based on their chain lengths. They have low thermal and electrical conductivity and low chemical reactivity and therefore are ideally suited for microactuator applications. Paraffins are stable up to 250oC, therefore no boiling occurs even at high temperatures. Further liquid paraffins are non-polar and hence do not mix with the polar liquids such as water which are commonly used in microfluidic applications. They undergo phase change from solid to liquid which is accompanied by volumetric expansion. Differences in volume between phases of 5-40% are observed for various paraffins [152]. Paraffins are also stable through numerous phase change cycles and their properties remain constant over long periods of time. Phase change is one of the preferred methods of actuation because of high displacements and forces obtained. Some of the early studies were focused on developing mini paraffin actuators for endoscopic surgery [153]. These studies led to the development of a 130 mm x 2 mm actuator with a 90 mm x 2 mm paraffin chamber. Volumetric expansion of 14 – 16% were achieved with pressure generation of upto 20 Mpa and stroke length of 9-
Materials and Microfabrication Processes
77
10 mm. Non traditional techniques such as screen printing have also been used to pattern paraffin in reservoirs 1 mm in radius and 0.5 mm in height [154]. Several such reservoirs arranged in series functioned as a peristaltic pump generating a flow rate of 74 nl/min. The main problems with this configuration was reported to be long response times (in seconds) and leakage of paraffin due to insufficient adhesion between the actuation membrane and the paraffin reservoirs. Although this early demonstration showed the promise of paraffin as an actuation material, integration of phase change materials into microsystems has been challenging [155] due to the fact that the material has to be introduced during post fabrication stage. This restricts the size of the actuator to upwards of 1 mm or more. Development of microfabrication techniques for paraffins in their solid form and its compatibility with polymer processing techniques has enabled ease of integration and development of miniature high force actuators in the microscale [105]. Paraffins have high vapor pressures and consequently evaporation is the preferred technique for thin film deposition. Deposition conditions of 5 µTorr chamber pressure and 150oC evaporation temperature provided a deposition rate of 100 nm/min. Adhesion of the deposited film to the substrate was enhanced by silanation and oxygen plasma treatment of the substrate prior to deposition. Surface roughness of 42 nm and less have been measured [152]. Pattering of this paraffin thin film followed an elaborate procedure involving 5 steps due to its temperature sensitivity and dissolution in solvents such as acetone and developer that are used in photolithography. Microactuators consisting of a microheater with a patterned micropatch of paraffin enclosed in a parylene sealed chamber have been developed [156]. Deflection of 2.7 µm at 150 mW was observed. Actuators were assembled into a valve format and a 6bit microflow controller was developed which had flow control from 0.1 – 2 sccm using 5 V, 50 mW power [156]. Another interesting use of Paraffin has been as a modifiable, mobile physical separator between microfluidic channels [157]. In this manifestation, the paraffin is introduced into one arm (stem channel) of a T junction microchannel network as shown in Figure 2.16. The other two arms are the inlet and the outlet. The stem channel is instrumented with several microheaters at various locations. The melted paraffin is introduced into the stem channel and flows by capillary action. The paraffin solidifies as it flows in the channel losing heat to the surroundings. Electronic control of the heaters in the stem channel allows precise positioning of the paraffin interface. The paraffin is initially allowed to flow and solidify at the T junction and physically separate the inlet and outlet arm. The valve is
78
Noori, Upadhyaya, and Selvaganapathy
latched in this position and blocks pressures as high as 250 psi without additional energy input. Subsequently the valve can be opened by activating the heaters, melting the paraffin and applying a pulse of vaccum at the stem channel interconnect. The duration of the pulse and the temperature of the paraffin determine the movement of the paraffin plug.
Fig. 2.16 Schematic of the phase change valve operation. (i) Loading wax by actuating inlet port heater. (ii) Closing valve by actuating the inlet port and stem channel heaters with pressure at inlet port. (iii) Opening valve by actuating the stem channel and intersection heaters with vacuum at the inlet port. (Reproduced with permission from [157] © (2004) American Chemical Society)
2.5.2.8 Biodegradable materials
Biodegradable materials have traditionally been used as scaffolds for tissue engineering applications [158], controlled release mechanism for drug delivery [159], and in sutures [160]. Commonly used biodegradable polymers are polyglycolic acid (PGA), polylactic acid (PLDA), polycaprolactone (PCL) and poly (lactide co glycolide) (PLGA). The degradation times of these polymers range from 5 months to greater than 24 months [161]. These polymers have low toxicity and minimal immune response, tunable degradation rates, and excellent mechanical properties suited for tissue engineering applications. The degradation of these polymers is by hydrolysis of unstable linkage groups in the backbone yielding degradation products that are capable of re-absorption by the body through metabolic pathways. For example, the degradation product of PGA is glycolic acid, which is a natural metabolite. Suitable unstable groups include esters, anhydrides, orthoesters, and amides. The structures of these polymers allow tailoring of their degradation rates. Synthesis of these polymers is carried out from their stable dimeric version. Figure 2.17 shows the chemical structure and synthesis of most common of the biodegradable polymers. The dimer is cyclic and hence
Materials and Microfabrication Processes
79
stabilized. In the presence of heat and an appropriate catalyst, it undergoes ring opening polymerization and forms linear chains. They take the shape of the mold in which the reaction occurs. Other techniques like injection molding, extrusion have been used for shaping these polymers. An excellent review of the material properties of various classes of biodegradable materials has been recently published [162].
Fig. 2.17 Structure and synthesis of common biodegradable materials.
A wide variety of microfabrication methods including imprinting or hot embossing [161], soft lithography [163], direct writing [163], stereolithography [164], and laser micromachining [165] have also been used for structuring biodegradable materials. Photocurable biodegradable versions of PCL have been developed for setereolithographic applications [166].
80
Noori, Upadhyaya, and Selvaganapathy
Microfabrication of these photopolymers using lithography has also been accomplished [167]. Compression molding combined with low temperature bonding of layers has been used to demonstrate the feasibility of creating microstructures, specifically microchannels, using biodegradable materials [168]. The application of these devices is in tissue engineering of microfluidic vascular networks, which then form the scaffold for seeding and growth of cells to form tissues with pre-existing blood supply networks. Biodegradable materials are attractive since they provide the initial structural framework for growth and then wither away when the tissue is able to sustain itself. The masters for molding containing the fluidic network are created out of PDMS using soft lithography techniques. PDMS posts are then attached to the reservoirs in the mold. Another PDMS master with larger holes that are concentric with the attached post is assembled on the first master. PLGA pellets are placed between these two layers and heated to 110oC along with application of compressive force of 100-500 lbs. The PLGA becomes a viscous melt and conforms to the microstructure of inside of the mold cavity. Film thickness, 100-500 µm, can be obtained by varying the compression time, temperature and applied force. An unpatterned PLGA film is used to close the microchannel structure using low temperature fusion bonding at 60oC for 30-60 min. Bonding at temperature slightly higher than glass transition temperature maintains structural integrity while still allowing interdiffusion in the areas of contact providing higher bond strength [168]. Multichannel stacks could be created this way leading to 3-D microchannel network as shown in Figure 2.18. A similar method was used to mold PCL material into microchambers for drug delivery [169].
Materials and Microfabrication Processes
81
Fig. 2.18 Microfluidic networks created in biodegradable materials. A) Complex branching of network of microchannels (Scale bar 500 µm); B) High Resolution lines (Scale bar 2 µm); C) Nominal microfluidic channels (Scale bar 50 µm); D) High resolution microchannels (Scale bar 2 µm); Reproduced with permission from [168] © 2004 Wiley.
Another interesting technique for creation of a microfluidic channel network for tissues has been the use of cell seeded alginate as the structural material that is casted on a microfabricated mold [170]. In this study calcium alginate is used as the tissue scaffold due to several favorable properties such as its appropriateness for long-term culture and formation of functional tissue, compatibility with macroscale molding to form cellseeded structures with physiological geometries and its ease of chemical modification to enable the presentation of biospecific extracellular ligands [170]. In this process PDMS masters with the appropriate microchannel network features were created using SU-8 lithographic techniques described above. Aluminum molding jigs were created and the master was placed in it. Subsequently cell-seeded alginate (4% (w/v) calcium alginate) was injected and cross-linked. Multiple layers of the molded alginate were sealed together by application of pressure by the aluminum jig. Replication of 100 µm channels have been achieved using this technique [170] with high cellular viabilities. The degree of viability after fabrication was sensitive to the magnitude of shear imposed during the dispersal and injection of the cells in the alginate solution.
82
Noori, Upadhyaya, and Selvaganapathy
2.6 Conclusion There has been growing interest and effort over the past 15 years towards development of miniaturized and automated devices for chemical and biochemical processes and analysis in areas as diverse as medical diagnostics, drug discovery, specialty chemical production, cell handling, and analytical laboratory applications. A significant majority of microfluidic devices are currently fabricated in silicon and glass primarily due to established fabrication techniques that have historically been developed for these materials. A further advantage is that most of the biochemical reactions have been characterized and standardized in glass. The use of polymers in microfluidics has been recent but is rapidly growing. A number of microfabrication methods, both conventional and non-conventional, have been developed to encompass a wide selection of polymers available. The attractiveness of polymers is due to their low cost, variety, solvent, and biocompatibility and tailorable surface properties. Individual components and devices such as microchannel networks, pumps, and valves have been successfully demonstrated using silicon, glass, and polymers. The focus of future research is in three main directions: 1) Development, modification, and optimization of existing microfabrication technologies for low-cost high-volume production of microfluidic devices, 2) Development of new materials and associated microfabrication techniques tailored for specific applications in microfluidics and 3) Development of integrated systems and packaging processes involving multiple materials and fabrication technologies for complex functionality. Combined together, these developments have the potential to radically change biochemical analysis, drug delivery, drug discovery, and tissue engineering. A number of these devices have already permeated into use by non-specialists in microfabrication, demonstrating their long term potential and success.
References 1. Campbell SA (1996) The science and engineering of microelectronic fabrication. Oxford University Press, Oxford. 2. Rajurkar K, Madou M (2007) Micromachining: International research and development. Springer, Netherlands, pp. 53-87. 3. Dittrich PS, Manz A (2006) Lab-on-a-chip: microfluidics in drug discovery. Nature Reviews Drug Discovery 5:210-218. 4. Selvaganapathy P, Carlen ET, Mastrangelo CH (2003) Recent progress in microfluidic devices for nucleic acid and antibody assays. IN: Proc. of the IEEE 91:954- 975.
Materials and Microfabrication Processes
83
5. Terry SC, Jerman JH, Angell JB (1979) A Gas Chromatographic Air Analyzer Fabricated on a silicon Wafer. Trans Electron Dev 12:1880–1886. 6. Bassous E, Taub HH, Kuhn L (1977) Ink jet printing nozzle arrays etched in silicon. Applied Physics Letters 31:135-137. 7. Bassous E, Baran EF (1978) The Fabrication of High Precision Nozzles by the Anisotropic Etching of (100) Silicon. Journal of the Electrochemical Society 125:1321-1327. 8. Petersen KE (1979) Fabrication of an integrated planar silicon ink-jet structure. IEEE Trans. Electron Devices 26:1918-1920. 9. Manz A, Graber N, Widmer HM (1990) Miniaturized total analysis systems: A novel concept for chemical sensors. Sensors and Actuators B: Chemical 1:244-248. 10. Burns MA, Mastrangelo CH, Sammarco TS, Man FP, Webster JR, Johnson BN, Foerster B, Jones D, Fields Y, Kaiser AR, Burke DT (1995) Microfabricated structures for integrated DNA analysis. Proc. of the National Academy of Sciences 93:5556-5561. 11. Reston RR, Kolesar ES (1994) Silicon-micromachined gas chromatography system used to separate and detect ammonia and nitrogen dioxide. I. Design, fabrication, and integration of the gas chromatography system. Journal of Microelectromechanical Systems 3:134-146. 12. Jensen KF (1999) Microchemical systems: status, challenges, and opportunities. AIChEJournal 45:2051-2054. 13. Verpoorte E (2002) Microfluidic chips for clinical and forensic analysis. Electrophoresis 23: 677- 712. 14. Gardeniers H, Van den Berg A (2004) Micro- and nanofluidic devices for environmental and biomedical applications. International Journal of Environmental Analytical Chemistry 84:809-819. 15. Reyes DR, Iossifidis D, Auroux PA, Manz A (2002) Micro Total Analysis Systems. 1. Introduction, Theory, and Technology. Analytical Chemistry 74:2623-2636 16. Gardner JW, Varadan VK, Awadelkarim OO (2001) Microsensors, MEMS, and Smart Devices. Wiley, New York. 17. Bhushan B (2004) Handbook of Nanotechnology. Springer, Berlin. 18. Petersen KE (1982) Silicon as a Mechanical Material. Proceedings of the IEEE 70:420-457. 19. Maluf N, Williams K (2004) An Introduction to Microelectromechanical Systems Engineering. Artech House Inc, Boston. 20. Korvink JG, Paul O (2006) MEMS: A Practical Guide to Design, Analysis, and Applications. William Andrew Publishing, Norwich 21. Bean KE (1978) Anisotropic Etching of Silicon. IEEE Transactions on Electron Devices 25: 1185- 1193. 22. Gregory TA, Kovacs A, Maluf NI, Petersen KE (1998) Bulk Micromachining of Silicon. In: Proceedings of the IEEE 86: 1536-1551. 23. Chen J, Wise KD, Hetke JF, Bledsoe SC (1997) A multichannel neural probe for selective chemical delivery at the cellular level. IEEE Transactions on Biomedical Engineering 44:760-769.
84
Noori, Upadhyaya, and Selvaganapathy
24. Laermer F, Urban A (2003) Challenges, developments and applications of silicon deep reactive ion etching. Microelectronic Engineering 67–68:349–355 25. Li S, Day JC, Park JJ, Cadou CP, Ghodssi R (2007) A fast-response microfluidic gas concentrating device for environmental sensing. Sensors and Actuators A 136:69–79. 26. Bustillo JM, Howe RT, Muller RS (1998) Surface micromachining for microelectromechanical systems. Proceedings of the IEEE 86: 1552-1574. 27. Lee KB, Lin L (2003) Surface micromachined glass and polysilicon microchannels using MUMPS for BioMEMS applications. Sensors and Actuators A 111:44–50. 28. Lin L, Pisano AP (1999) Silicon-processed microneedles. Journal of Microelectromechanical Systems 8:78-84. 29. Feingold RM (1979) Method of making a plate having a pattern of microchannel US Patent 4153855. 30. Kopf DJ (1984) Pipete puller. US Patent 4530712. 31. Saito T (1988) Microchannel plate and a method for manufacturing the same. US Patent 4780395. 32. van Lintel HTG, van de Pol FCM, Bouwstra S (1988) A piezoelectric micropump based on micromachining of silicon. Sensors Actuators B 15:153–167. 33. Harrison DJ, Fluri K, Seiler K, Fan ZH, Effenhauser C. S, Manz A (1993) Micromachining a Miniaturized Capillary Electrophoresis-Based Chemical Analysis System on a Chip. Science 261:895-897. 34. McClain MA, Culbertson CT, Jacobson SC, Ramsey JM (2001) Flow cytometry of Escherichia coli on microfluidic devices. Anal Chem 73:53345338. 35. Moore AW, Jacobson SC, Ramsey JM (1995) Microchip Separations of Neutral Species via Micellar Electrokinetic Capillary Chromatograph. Anal Chem 67:4184-4189. 36. Khandurina J, McKnight TE, Jacobson SC, Waters LC, Foote RS, Ramsey JM (2000) Integrated System for Rapid PCR-Based DNA Analysis in Microfluidic Devices. Anal Chem 72:2995-3000. 37. Xue Q, Foret F, Dunayevskiy YM, Zavracky PM, McGruer NE, Karger BL. (1997) Multichannel Microchip Electrospray Mass Spectrometry. Anal Chem 69:426-430. 38. Lagally ET, Simpson PC, Mathies RA (2000) Monolithic integrated microfluidic DNA amplification and capillary electrophoresis analysis system. Sensors and Actuators B 63:138-146. 39. Tso ST, Pask JA (1982) Reaction of Glasses with Hydrofluoric Acid Solution. Journal of American Ceramic Society 65:360-363. 40. Spierings GACM (1993) Wet chemical etching of silicate glasses in hydrofluoric acid based solutions: Review. Journal of Materials Science 28:6261-6273. 41. Rangsten P, Hedlund C, Katardjiev IV Backlund Y (1998) Etch rates of crystallographic planes in Z -cut quartz—experiments and simulation. J Micromech Microeng 8:1–6.
Materials and Microfabrication Processes
85
42. Ueda T, Kohsaka F, Yamazaki D (1985) Quartz crystal micromechanical devices. In: Proc. 3rd Int. Conf Solid-State Sensors and Actuators (Transducers ' 85) Technical digest pp 113-117. 43. Jacodine RJ (1962) Use of modified free energy theorems to predict equilibrium growing and etching shapes. J Appl Phys 33:2643-2647. 44. Ueda T, Kohsaka F, Lino T, Yamazaki D (1987) Theory to predict etching shapes in quartz and application to design devices. Trans Soc Instrum Control Eng 23:1233-1238. 45. Grosse A, Grewe M, Fouckhardt H, (2000) Deep wet etching of fused silica glass for hollow capillary optical leaky waveguides in microfluidic devices. J Micromech Microeng 11:257–262. 46. Stjernstrom M, Roeraade J (1998) Method for fabrication of microfluidic systems in glass. J Micromech Microeng 8:33–38. 47. Ronggui S, Righini GC (1991) Characterization of reactive ion etching of glass and its applications in integrated optics. Journal of Vacuum Science & Technology A 9:2709-2712. 48. Li X, Abe T, Esashi M, (2000) Deep reactive ion etching of Pyrex glass using SF6 plasma. Sensors and Actuators A 87:139-145. 49. Abe T, Esashi M (1999) One-chip multi-channel quartz crystal microbalance (QCM) fabricated by Deep RIE. In: Proc. of Transducers '99, pp 1246-1249. 50. Dietrich TR, Ehrfeld W, Lacher M, Kramer M, Speit B (1996) Fabrication technologies for microsystems utilizing photoetchable glass. Microelectron Eng 30:497-504. 51. Bings NH, Wag C, Skinner CD, Colyer CL, Thibault P, Harrison DJ (1999) Microfluidic devices connected to fused silica capillaries with minimal dead volume. Anal Chem 71:3292-3296. 52. Ngoi BKA, Sreejith PS (2000) Ductile regime finish machining—a review. Int J Adv Manuf Technol 16: 547–550. 53. Ligthart HJ, Slikkerveer PJ, IntVeld FH, Swinkels PHW, Zonneveld MH (1996) Glass and glass machining in Zeus panels. Philips Journal of Research 50:475-499. 54. Buijs M (1994) Erosion of glass as modeled by indentation theory. J Am Ceram Soc 77:1676–1678. 55. Wensink H, Jansen HV, Berenschot JW, Elwenspoek MC (2000) Mask materials for powder blasting. J Micromech Microeng 10:175–180. 56. Slikkerveer PJ, Bouten PCP, de Haas FCM, (2000) High quality mechanical etching of brittle materials by powder blasting. Sensors Actuators 85:296– 303. 57. Solignac D (2003) Glass microchips for biochemical analysis: Technologies and applicaitions, Ph.D Thesis, EPFL Lausanne. 58. Schlautmann S, Wensink H, Schasfoort R, Elwenspoek M, van den Berg A (2001) Powder blasting technology as a alternative tool for microfabrication of capillary electrophoresis chips with integrated conductivity sensors. Journal of Micromechanics and Microengineering 11:386-389. 59. Wensink H, (2002) Ph.D. Thesis, University of Twente, Fabrication of microstructures by powder blasting. Enschede, Netherlands.
86
Noori, Upadhyaya, and Selvaganapathy
60. Basak I, Ghosh A (1996) Mechanism of Spark Generation during Electrochemical Discharge Machining: A Theoretical Model and Experimental Verification. Journal of Material Processing Technology 62: 46-53. 61. Zheng ZP, Cheng WH, Huang FY, Yan BH (2007) 3D microstructuring of Pyrex glass using the electrochemical discharge machining process. J Micromech Microeng 17:960–966. 62. Schimdt MA (1998) Wafer-to-Wafer Bonding for Microstructure Formation. In: Proceedings of the IEEE 86:1575-1585. 63. Wallis G, Pomerantz DI (1969) Field assisted glass-metal sealing. Journal of Applied Physics 40:3946–3949. 64. Acero MC, Plaza JA, Esteve J, Carmona M, Marco S, Samitier J (1997) Design of a modular micropump based on anodic bonding. J Micromech Microeng 7:179–182. 65. Lee SW, Sim WY, Yang SS (1999) Fabrication of a micro syringe. Proc of Transducers, Sendai, Japan 1368–1371. 66. Hedenstierna N, Habibi S, Nilsen SM, Kvisteroy T, Jensen GU (2001) Bulk micromachined angular rate sensor based on the “butterfly” gyro structure. In: Proceedings of MEMS Interlaken, Switzerland, 178–181. 67. Chavan AV, Wise KD, (2002) A monolithic fully-integrated vacuum sealed CMOS pressure sensor. IEEE Trans Electron Devices 49:164–169. 68. Nitzsche P, Lange K, Schmidt B, Grigull S, Kreissig U, Thomas B, Herzog (1998) Ion drift processes in Pyrex-type alkaliborosilicate glass during anodic bonding. J Electrochem Soc 145:1755–1762. 69. Kanda Y, Matsuda K, Murayama C, Ugaya J, (1990) The mechanism of fieldassisted silicon-glass bonding. Sens Actuators A 21-23:939–943. 70. Albaugh KB, (1991) Electrode phenomena during anodic bonding of silicon to borosilicate glass. J Electrochem Soc 138:3089–3094. 71. Albaugh KB, Cade PE Rasmussen DH. (1988) Mechanism of anodic bonding of silicon to pyrex glass. Technical Digest IEEE Solid State Sensor and Actuator Workshop, 109-110. 72. Wallis G. (1970) Direct-Current Polarisation during Field-Assisted GlassMetal Sealing. Journal of The American Ceramic Society 53:563-67. 73. Niklaus F, Enoksson P, Kalvesten E, Stemme G (2001) Low-temperature full wafer adhesive bonding. J Micromech Microeng 11:100–107 74. Alvino WM (1995) Plastics For Electronics: Materials, Properties, and Design. New York: McGraw-Hill. 75. Booth E, Hunt CE (1995) Low temperature adhesion bonding methods Proc. Semiconductor Wafer Bonding: Science, Technology and Applications. J Electrochem Soc 95-97:201–211. 76. Wu H, Bo Huang B, Zare RN (2005) Construction of microfluidic chips using polydimethylsiloxane for adhesive bonding. Lab Chip 5:1393 – 1398. 77. Nicholson W (1997) The Chemistry of Polymers, 2nd Edition, Royal Society of Chemistry, Cambridge, UK. 78. Becker H, Gartner C (2000) Polymer microfabrication methods for microfluidic analytical applications. Electrophoresis 21:12-26.
Materials and Microfabrication Processes
87
79. Heckele M, Schomburg W K (2004) Review on micro molding of thermoplastic polymers. J Micromech Microeng 14:R1–R14. 80. Hanemann T, Heckele M, Piotter V (2000) Current status of micromolding technology. Polym News 25:224–229. 81. Weber L, Ehrfeld W, Freimuth H, Lacher M, Lehr H, Pech B (1996) Micromolding—a powerful tool for the large scale production of precise microstructures. In: Proc SPIE 2879:156–167. 82. Despa M S, Kelly K W, Collier J R (1999) Injection molding of polymeric LIGA HARMs. Microsyst Technol 6:60–66. 83. Piotter V, Bauer W, Benzler T, Emde A (2001) Injection molding of components for Microsystems. Microsyst Technol 7:99–102. 84. Larsson O, Ohman O, Billman A, Lundbladh L, Lindell C, Palmskog G (1997) Silicon based replication technology of 3D-microstructures by conventional CD-injection molding techniques. In: Proc Int Conf on Solid-State Sensors and Actuators 1415–1418. 85. Jian L, Desta Y M, Goettert J (2001) Multilevel microstructures and mold inserts fabricated with planar and oblique x-ray lithography of SU-8 negative photoresist. In: Proc. SPIE 4557:69–76. 86. Ruprecht R, Hanemann T, Piotter V, Hausselt J (1995) Injection molding of LIGA and LIGA-similar microstructures using filled and unfilled thermoplastics. In: Proc SPIE 2639:146–157. 87. Sethu P, Mastrangelo C H (2004) Cast epoxy-based microfluidic systems and their application in biotechnology. Sens Actuators B: Chem 98(2–3):337–346. 88. Xia, G. M. Whitesides G M (1998) Soft Lithography. Angew Chem Int Ed 37:550-575. 89. Hull C W (1986) Apparatus for production of three dimensional objects by stereolithography. US patent 4,575,330. 90. Takagi T, Nakajima N (1993) Photoforming applied to fine machining. In: Proc Of 4th Int Sym On Micromachinig and Human Science 173-178. 91. Ikuta K, Hirowatari K (1993) Real 3-D microfabrication using stereolithography and metal molding. In: Proc Of IEEE workshop on MEMS 42-47. 92. Zhang X, Jiang X N, Sun C (1998) Microstereolithography for MEMS. In: Proc of IEEE Conf On MEMS 3-9. 93. Ikuta K, Ogata T, Tsubio M, Kojima S (1996) Development of mass productive micro stereolithography. In: Proc of IEEE Conf On MEMS 301-306. 94. Bertsch A, Zissi S, Jezequel J Y, Crobel S, Andre J C (1997) Microstereolithography using a liquid crystal display as a dynamic mask generator. Micro Tech 3:42-47. 95. Bertsch A, Lorenz H, Renaud P (1998) Combining microstereolithography and thick resist UV lithography for 3D microfabrication. In: Proc of IEEE Conf On MEMS, 18-23. 96. Bertsch A, Bernhard P, Renaud P (2001) Microstereolithography: Concepts and Applications. In: Proc of IEEE Conf on Emerging Technologies and Factory Automation 289-298.
88
Noori, Upadhyaya, and Selvaganapathy
97. Lin AW, Wong CP (1992) Encapsulant for nonhermetic multichip packaging applications. IEEE Trans on Components, Packaging, and Manufacturing Technology 15:510-518. 98. Olson R (1989) Parylene conformal coatings and their applications for electronics. In: Proc in 19th Electrical Electronics Insulation Conference 272-273. 99. Selbrede SC, Zucker ML (1997) Characterization of parylene-n thin films for low-k vlsi applications. In: Low-Dielectric Constant Materials III Materials society symposium Proc 476:219-224. 100. Flaherty M (1995) Conformal polymer film protects circuits, stabilizes solder joints. In: Proc of the Int. Electronics Packaging Conf 675-679. 101. Webster JR, Mastrangelo CH (1997) Large-volume integrated capillary electrophoresis stage fabricated using micromachining of plastics on silicon substrates. In: Proceedings of the 1997 International Conference on Solid- State Sensors and Actuators (Tranducers ’97) 503–506. 102. Man PF, Jones DK, Mastrangelo CH (1997) Microfluidic plastic capillaries on silicon substrates: a new inexpensive technology for bioanalysis chips. In: Proc of the IEEE Micro Electro Mechanical Systems (MEMS), Nagoya, Japan 311–316. 103. Wang X, Lin Q, Tai YC (1999) Parylene micro check valve. In: Proc of the IEEE Micro Electro Mechanical Systems (MEMS) 177–182. 104. Rich CA, Wise KD (1999) An 8-bit microflow controller using pneumatically actuated microvalves. In: Proc of the IEEE Micro Electro Mechanical Systems (MEMS) 130–134. 105. Carlen ET, Mastrangelo CH (2002) Electrothermally activated paraffin microactuators. J MEMS 11:165-174. 106. Yang X, Yang JM, Wang XQ, Meng E, Tai YC, Ho CM (1998) Micromachined membrane particle filters. In: Proceedings of the IEEE Micro Electro Mechanical Systems (MEMS) 137–142. 107. Lahann J, Klee D, Hocker H (1998) Chemical vapour deposition polymerization of substituted [2.2] paracyclophanes Macromol Rapid Commun 19:441– 444. 108. Webster JR, Burke DT, Burns MA, Mastrangelo CH (1998) An inexpensive plastic technology for microfabricated capillary electrophoresis chips. In: Proc of the µTas ’98 Workshop 249–252. 109. Tacito RD, Steinbruchel C (1996) Fine-line patterning of parylene-n by reactive ion etching for application as an interlayer dielectric. J Electrochemical Soc 143:1973–1977. 110. Bell TE, Wise KD, Anderson DJ (1997) Flexible micromachined electrode array for a cochlear prosthesis. In: International Conference on Solid-State Sensors and Actuators Proceedings Vol. 2 (Tranducers ’97) 1315–1318. 111. Lee KY, LaBianca N, Zolgharnain S, Rishton SA, Gelorme JD, Shaw JM, Chang THP, (1995) Micromachining applications of a high resolution ultrathick photoresist. J Vac Sci Technol Vol. B 13: 3012–3016. 112. Lorenz H, Despont M, Fahrni N, LaBianca , Renaud P, Vettiger P, (1997) SU-8: a low-cost negative resist for MEMS. J Micromech Microeng 7:121– 124.
Materials and Microfabrication Processes
89
113. Lorenz H, Despont M, Fahrni N, Brugger J, Vettiger P, Renaud P, (1998) High-aspect-ratio, ultra thick, negative-tone near-UV photoresist and its applications for MEMS. Sensors and Actuators A 64:33-39. 114. Jian L, Desta YM, Goettert J, Bednarzik M, Loechel B, Yoonyoung J, Aigeldinger G, Singh V, Ahrens G, Gruetzner G, Ruhmann R, Degen R, (2003) SU-8 based deep x-ray lithography/LIGA. In: Micromachining and Microfabrication Process Technology VIII, Proceedings of SPIE 4979:394401. 115. Guerin LJ, Bossel M, Demierre M, Calmes S, Renaud P, (1997) Simple and low cost fabrication of embedded microchannels by using a new thick film photoplastic. In: Proc. of Transducers’ 97, Chicago USA 1419-1422. 116. Tay FEH, van Kan JA, Watt F, Choong WO, (2001) A novel micromachining method for the fabrication of thick-film SU-8 embedded microchannels. J Micromech Microeng 11:27–32. 117. Heuschkel MO, Guerin L, Buisson B, Bertrand D, Renaud P, (1998) Buried microchannels in photopolymer for delivering of solutions to neurons in a network. Sensors and Actuators B 48:356–361. 118. Brook M, (2000) Silicon in Organic, Organometallic and Polymer Chemistry, John Wiley and Sons, New York. 119. Effenhauser CS, Bruin GJ, Paulus A, Ehrat M (1997) Integrated Capillary Electrophoresis on Flexible Silicone Microdevices: Analysis of DNA Restriction Fragments and Detection of Single DNA Molecules on Microchips. Anal Chem 69:3451-3457. 120. Duffy DC, McDonald JC, Schueller OJ, Whitesides GM (1998) Rapid Prototyping of Microfluidic Systems in Poly(dimethylsiloxane). Anal Chem 70:4974-4984. 121. Hillborg H, Ankner JF, Gedde UW, Smith GD, Yasuda HK, Wikstrom K (2000) Crosslinked polydimethylsiloxane exposed to oxygen plasma studied by neutron reflectometry and other surface specific techniques. Polymer 41:6851-6863. 122. Efimenko K, Wallace WE, Genzer J (2002) Surface Modification of Sylgard184 Poly(dimethyl siloxane) Networks by Ultraviolet and Ultraviolet/Ozone Treatment. J Colloid Interface Sci 254:306-315. 123. Hillborg H, Gedde UW (1998) Hydrophobicity recovery of polydimethylsiloxane after exposure to corona discharges. Polymer Science 39:1991-1998. 124. Unger MA., Chou HP, Thorsen T, Scherer A, Quake SR (2000) Monolithic Microfabricated Valves and Pumps by Multilayer Soft Lithography. Science 288:113-116. 125. Goll C, Bacher W, Bustgens B, Maas D, Menz W, Schomburg WK (1996) Microvalves with bistable buckled polymer diaphragms. J Micromech Microeng 6:77-79. 126. Metz S, Holzer R, Renaud P (2001) Polyimide-based microfluidic devices. Lab on a Chip 1:29–34. 127. Hiramoto H (1990) Photosensitive Polyimides. In: Advanced Electronic Packaging Materials Proceedings, MRS 167:87-97.
90
Noori, Upadhyaya, and Selvaganapathy
128. Glasgow IK, Beebe DJ, White VE (1999) Design rules for polyimide solvent bonding. Sensors and materials 11:269-278. 129. Dusek K (1993) Advances in Polymer Science; Responsive Gels: Volume Transitions II., Springer-Verlag: Berlin 130. Tanaka T, Fillmore DJ (1979) Kinetics of swelling of gels. J Chem Phys 70:1214-1218. 131. Tanaka T, Fillmore D, Sun ST, Nishio I, Swislow G, Shah A (1980) Phase Transitions in Ionic Gels. Phys Rev Lett 45:1636–1639. 132. Shiga T (1997) Deformation and Viscoelastic Behavior of Polymer Gels in Electric Fields. Adv Polym Sci 134:131–163. 133. Kataoka K, Miyazaki H, Bunya M, Okano T, Sakurai Y (1998) Totally synthetic polymer gels responding to external glucose concentration: their preparation and application to on–off regulation of insulin release J Am Chem Soc 120:12694–12695. 134. Miyata T, Asami N, Uragami T (1999) A reversibly antigen-responsive hydrogel. Nature 399:766–769. 135. Gan DJ, Lyon LA (2001) Tunable Swelling Kinetics in Core-Shell Hydrogel Nanoparticles. J Am Chem Soc 123:7511–7517. 136. Pelton R (2000) Temperature-sensitive aqueous microgels. Adv in Colloid and Interface Science 85:1-33. 137. Chen G, Imanishi Y, Ito Y (1998) Photolithographic Synthesis of Hydrogels Macromolecules 31:4379-4381. 138. Mutlu S, Yu C, Svec F, Mastrangelo CH, Frechet JMJ, Gianchandani YB 2003) A thermally responsive polymer microvalve without mechanical parts photopatterned in a parylene channel. In: Proc of Conf on Solid State Sensors Actuators and Microsystems, 802-805. 139. Beebe DJ, Moore JS, Bauer JM, Yu Q, Liu RH, Devadoss C, Jo BH (2000) Functional hydrogel structures for autonomous flow control inside microfluidic channels. Nature 404:588-590. 140. Baldi A, Gu Y, Loftness PE, Siegel RA, Ziaie B (2003) A hydrogel-actuated environmentally sensitive microvalve for active flow control. JMEMS 12:613-621. 141. Peters EC, Svec F, Frechet JMJ (1997) The preparation of large diameter molded porous polymer monoliths and the control of pore structure homogenity. Chem Mater 9:1898-1902. 142. Brooks DW (1990) Basic aspects and recent developments in suspension polymerization. Macromol Chem Macromol Symp 35/36:121-125. 143. Svec F, Frechet JMJ (1995) Temperature, a simple and efficient tool for the control of pore size in macroporous polymers. Macromolecules 28:75807582. 144. Viklund C, Svec F, Frechet JMJ (1996) Monolithic molded porous materials with high flow characteristics for seperations, catalysis, or solid phase chemistry: control of porous properties during polymerization. Chem. Mater 8:744750. 145. Unger K (Eds), (1990) Packings and stationary phases in chromatographic techniques. New York: Dekker.
Materials and Microfabrication Processes
91
146. Wang Q C, Svec F, Frechet JMJ (1993) Macroporous polymer stationary phase rod as continous seperating medium for reverse phase chromatography. Anal Chem 65:2243-2248. 147. Svec F, Frechet JMJ (1995) Molded rods of polymer for preparative separation of biological products. Biotech Bioeng 48:476-480. 148. Svec F, Frechet JMJ (1995) Modified poly(glycidyl methacylate-co-ethylene dimethacrylate) continous rod columns for preparative-scale ion-exchange chromatography of proteins. J Chromatogr 702:89-95. 149. Petro M, Svec F, Frechet JMJ (1996) Immobilization of trypsin on molded macroporous poly(glycidyl methacrylate - co - ethylene dimethacrylate) rods and use of conjugates are bioreactors for affinity chromatography. Biotech Bioeng 49:355-363. 150. Wang Q C, Svec F, Frechet J M J (1995) Hydrophilization of porous polystyrene based continuous rod columns. Anal Chem 67:670-674. 151. Mutlu S, Yu C, Selvaganapathy P, Svec F, Mastrangelo CH, Frechet JMJ, (2002) Micromachined porous polymer for bubble free electroosmotic pump. In: Proc of IEEE Conf on MEMS 19 -23. 152. Carlen ET (2001) Electrothermally actuated polymer microvalves. Ph.D Thesis, University of Michigan. 153. Kabei N, Kosuda M, Kagamibuchi H, Tashiro R, Mizuno H, Ueda Y, Tsuchiya K (1997) A Thermal-Expansion-Type Microactuator with paraffin as the Expansive Material. JSME Int Journal C 40:736-742. 154. Boden R, Lehto M, Simu U, Thornell G, Hjort K, Schweitz JA (2006) A polymeric paraffin actuated high-pressure micropump. Sensors and Actuators A 127:88–93. 155. Bergstrom PL, Ji J, Liu YL, Kaviany M, Wise KD (1995) Thermally driven phase change microactuation. J MEMS 4:10-17. 156. Carlen, ET, Mastrangelo CH (2000) Paraffin actuated surface micromachined valves. IEEE Conf on MEMS 381-385. 157. Pal R, Yang M, Johnson BN, Burke DT, Burns MA (2004) Phase change microvalve for integrated devices. Anal Chem 76:3740-3748. 158. Freed LE, Novakovic GV, Biron RJ, Eagles DB, Lesnoy DC, Barlow SK, Langer R (1994) Biodegradable polymer scaffold for tissue engineering. Nature Biotechnology 12:689- 693. 159. Langer R (1990) New methods of drug delivery. Science 249:1527– 1533. 160. Frazza EJ, Schimdt EE (1971) A new absorbable suture. J Biomed Mater Res Symp 1:43-58. 161. Lu Y, Chen SC (2004) Micro and nano-fabrication of biodegradable polymers for drug delivery. Adv Drug Delivery Rev 56:1621– 1633. 162. Gunatillake PA, Adhikari R (2003) Biodegradable synthetic polymers for tissue engineering. European cells and materials 5:1-16. 163. Vozzi G, Flaim C, Ahluwalia A, Bhatia S (2003) Fabrication of PLGA scaffolds using soft lithography and microsyringe deposition. Biomaterials 24:2533–2540.
92
Noori, Upadhyaya, and Selvaganapathy
164. Cooke MN, Fisher JP, Dean D, Rimnac C, Mikos AG (2002) Use of stereolithography to manufacture critical-sized 3D biodegradable scaffolds for bone ingrowth. J Biomed Mater Res Part B: Appl Biomater 64B:65– 69. 165. Kancharla V, Chen SC (2002) Fabrication of biodegradable polymeric microdevices using laser micromachining. Biomed Microdev 4:105–109. 166. Matsuda T, Mizutani M, Arnold SC (2000) Molecular design of photocurable liquid biodegradable copolymers. 1.Synthesis and photocuring characteristics. Macromolecules 33:795–800. 167. Leclerca E, Furukawab KS, Miyatab F, Sakaic Y, Ushidab T, Fujiic T (2004) Fabrication of microstructures in photosensitive biodegradable polymers for tissue engineering applications. Biomaterials 25:4683–4690. 168. King KR, Wang C, Kaazenpur M, Vacanti JP, Borenstein JT (2004) Biodegradable Microfluidics. Adv Mater 16:2007-2012. 169. Armani DK, Liu C (2000) Microfabrication technology for polycaprolactone, a biodegradable polymer. J Micromech Microeng 10:80-84. 170. Choi NK, Cabodi M, Held B, Gelghorn JP, Bonassar LJ, Stroock AD (2007) Microfluidic scaffolds for tissue engineering. Nature Materials 6:908-915.
Chapter 3 Interfacing Microfluidic Devices with the Macro World
Xuefeng Wang and Kashan A. Shaikh GE Global Research, 1 Research Circle, Niskayuna NY 12309
Keywords: microfluidic interface, chip-to-world, electrospray mass spectrometry, emitter, drug delivery, fluidic couplers
Abstract The universal requirement for all microfluidic devices is a robust fluidic interface between the device and the macro world. This interface consists of two main types of connections: 1) fluid delivery to the device from the macro world, and 2) output from the device to the external surroundings. Microfluidic interfaces must satisfy a number of different criteria including being reliable, mechanically robust, leak free, simple to assemble, and having minimal dead volume. Additionally, since the interface typically requires millimeter sized parts, fabrication processes deviate from the typical MEMS process. This chapter explores the current state of the art in worldto-chip and chip-to-world microfluidic interfaces. World-to-chip interface examples mainly consist of an integrated coupler that mates external capillary tubing to the device. Chip-to-world interface examples include interfacing the microfluidic device to other instruments or directly to biological subjects (e.g., humans).
94
Wang and Shaikh
3.1 Introduction Microfluidic systems have been widely studied in recent years for chemical detection, biological analysis, medical therapeutics, and pharmaceutical applications. The small volume and large surface to volume ratio enabled by this technology allows chemical and biological processes to be realized with less sample, lower cost, faster speed, and better accuracy than traditional laboratory processes. A microfluidic device typically consists of a combination of fluidic control and sensing components, such as channels, chambers, valves, pumps, mixers, heaters, and sensors. While selective integration of these components is largely application dependent, all devices require an interface between the microdevice and macro world. There are two main types of interconnections: the interface to microfluidic device from macro world for sample loading, and the interface from microdevice to external world for fluidic output. Though numerous passive and active microfluidic components have been demonstrated to date, the microfluidic interface remains an area that is relatively overlooked and under-studied. However, a reliable microfluidic interface is critical for any microfluidic product to be commercially viable.
3.2 Typical Requirements for Microfluidic Interfaces The solution for a successful microfluidic interface must be able to address several issues. First, it needs to be reliable, mechanically robust, and leak free. This allows fluidic samples to be introduced into or out of microfluidic chip securely and without contamination. Secondly, it needs to avoid introducing dead volume. Due to structural mismatches between the interface and microchannel, dead volume is easily created. Existence of dead volumes may reduce the efficiency of the microfluidic system and introduce sample contamination issues. Lastly, it needs to allow simple assembly and the flexibility to attach and detach with relative ease. Many microfluidic devices are designed for disposable use to avoid sample crosscontamination. This requires simple mechanisms to connect devices to external instruments in order to allow for efficient operation.
Interfacing Microfluidic Devices with the Macro World
95
3.3 Review of Microfluidic Interfaces 3.3.1 World-to-chip interfaces 3.3.1.1 Direct sample loading structures
The most commonly used interconnection between a microfluidic system and macroscopic world is the loading well [1]. Loading reservoirs, with one end open to the ambient environment and the other end connected to the microchannel network, can be created on various types of substrates using etching or machining techniques [2]. The size of loading reservoirs is typically in the millimeter scale to facilitate manual or automatic liquid loading using syringe needles or pipettes. The liquid loaded in the reservoir can then be drawn into the microchannels by capillary force, differential pressure, or electrokinetic force. 3.3.1.2 Interface adaptors
Interface adaptors are comprised of mechanical parts assembled together and are designed to quickly connect the external world to the microfluidic device. Unlike disposable microfluidic chips or cartridges, the interface adaptors are typical made using conventional machining tools and are intended for repeated use. Advantages include improved mechanical integrity and simplified assembly. Nittis et al. developed a mechanical assembly that provides an interface for coupling standard capillary tubing to microfluidic mixer chips [3]. As shown in Figure 3.1, the assembly measures 38x35x30 mm, and can be assembled in 5 min. When assembled, the recesses in the two main body halves form portholes to position the capillary tubing into the inlets of the microfluidic chip. The alignment of the microfluidic chip and the capillaries is achieved through the alignment guides on the precision positioner that holds the chip. The sealing membrane, which is made from a silicone elastomer, seals off the edges of the inlet and outlet holes to prevent leakage. In addition, the flexible membrane can deform to absorb any machining inaccuracy of the assembly. The assembly also has a glass observation window for optical visualization. In experiments, the interface module withstood flow rates up to 1.5 mL/min before fluid leakage was observed. Since the capillaries enter the chip directly, the dead volume at the fluidic interface is minimized. In addition, the interface does not use any adhesive material, which can cause contamination and clogging issues.
96
Wang and Shaikh
Fig. 3.1 Schematics of a mechanical assembly microfluidic interface de-veloped by Nittis et al. [3]: (a) assembled module; (b) exploded view (This figure is reproduced by permission of The Royal Society of Chemistry).
Fig. 3.2 Schematics of the microfluidic interface developed by Oh et al. [4]: (a) exploded view; (b) cross-section view of sample loading mode; (c) cross-section view of sample sealing mode (This figure is reproduced by permission of The Royal Society of Chemistry).
Oh et al. developed a microfluidic interfacing system with built-in valves for sample loading and sealing and demonstrated it on a multi-chamber po-
Interfacing Microfluidic Devices with the Macro World
97
lymerase chain reaction (PCR) chip [4]. The interface module consists of two plastic fittings and a plastic chip holder. In operation, a PCR chip is first inserted into the chip holder. Then the plastic fittings with rubber sealing sheets are assembled with the holder. PCR samples are loaded into the microchip directly through guide holes in the plastic fittings using pipettes. After the samples are loaded, the plastic fittings with the rubber sheets can be slid to seal the inlets and outlets of the chip without dead volume. Biocompatible Polydimethylsiloxane (PDMS) material is used to make the sealing sheet, to resist pressure building up inside the microfluidic channels. Leakage tests were conducted at elevated temperature (100oC, 30 min) on sealed chips and showed no detectable leakage. Real-time PCR assays were also conducted and yielded a 100% success rate with no contamination or leakage. 3.3.1.3 Micromachined fluidic couplers
Glass or plastic tubing is normally used in chemical analysis systems to transfer fluidic samples. Capillary tubing can be directly connected to a microchip using an adhesive bonding method. To improve the structural integrity, fitting holes can be created on glass or Si substrates using a drilling or etching method. Capillaries can be inserted in the fitting holes and be fixed by adhesives such as epoxy or crystal bond. For capillary electrophoresis applications, glass substrates are preferable since they can sustain the high electrical voltage needed in the tests. For glass chips, Bings et al. reported that fitting holes drilled by standard pointed drills have a conicalshaped bottoms that lead to a dead volume at the capillary and microfluidic system interface [5]. The dead volume is found to result in significant band broadening effect. Bings et al. demonstrated that a flat tipped drill bit could be used to flatten out the conical shaped bottom of the fitting hole to minimize the effective dead volume at the capillary-microchip interface. Li et al. demonstrated use of a PDMS interconnect piece for coupling glass or plastic tubing to microfluidic channels [6]. Holes are punched into a piece of PDMS pad using a glass capillary. The PDMS pad can then be bonded (permanently or reversibly) to microfluidic systems made from Si, glass, PDMS, or other materials. Permanent bonding can be achieved by oxygen plasma treating of bonding interfaces prior to bonding or by using epoxy glues. Glass or Teflon plastic tubing can be inserted into the holes in the PDMS connector. The interconnection can be further enforced by applying epoxy glues. The bonding strengths, leakage rates, and pull-out forces were characterized by Li et al. and it was found that the circularity
98
Wang and Shaikh
of the holes punched in the PDMS and the cleanliness of the bonding surfaces are critical to achieve high bonding strength.
Fig. 3.3 Schematic of PDMS interconnect developed by Li et al. [6] (This figure is reproduced by permission of the IEEE, ©2003 IEEE).
Due to geometric mismatches, tubing connections are known to introduce dead volumes into the microfluidic system. To alleviate this problem, Chiou et al. used a cast PDMS channel to bridge between a capillary and a microfluidic channel, as shown in Figure 3.4 [7]. When fabricating the fluidic interface, a capillary is first placed close to the opening of a microchannel. A thin metal wire is then threaded into the microchannel through the hollow capillary tubing. PDMS precursor is poured into a junction between the capillary and the microchannel. The PDMS is cured, locking the relative position of the capillary and the microchannel. Then the metal wire is removed from the PDMS, leaving a connection channel bridging between the capillary and microchannel. Due to surface tension, the PDMS connection channel has conical shaped ends with minimized dead volume. The length of the connection channel inside of the microfluidic system can be roughly controlled by pre-curing of the PDMS mixture and by varying the curing temperature. When needed, the capillary can be pulled out and replaced with another one of the same dimension. Leakage tests showed no leakage at the capillary-PDMS interface at pressures below 150 psi or flow rates below 50 µL/min. Pullout tests also showed the mechanical strength of the interface can be adjusted by varying the length of embedded capillary in the PDMS junction, with forces on the order of several Newtons obtainable.
Interfacing Microfluidic Devices with the Macro World
99
Fig. 3.4 Schematic of the minimal dead volume microfluidic interconnects from [7] (This figure is reproduced by permission of the IOP).
Han et al. developed a microfluidic system interface using the stereolithography (SLA) technique [8,9]. SLA is an additive fabrication method that uses a laser to fix photosensitive polymer structures layer by layer. It allows complex 3-dimensional microfluidic interfaces to be fabricated with 25 µm vertical (out-of-plane) resolution and 1 µm horizontal (in-plane) pattern resolution. Since the structures are created in a layer-by-layer fashion, integration of external subjects, such as electrical, mechanical, and optical components, into the SLA fabrication process is possible. Shown in Figure 3.5 is a cross-sectional view of the reported microfluidic interconnect and channel system. The one-step plug-in interconnects are made from SLA, with rubber O-rings assembled post fabrication. The interconnect chip can be aligned onto the glass channel chips using tooling holes or alignment marks. A low viscosity epoxy adhesive is then flowed into the gap between the interconnect and the glass chip through a dedicated adhesive via. The epoxy is then cured, to permanently bond the interconnect and the channel chips. To introduce the sample fluids, glass capillary tubing is plugged into the inlet and outlet ports in the interconnect chip. The first rubber O-ring between the interconnect layer and the glass chip is used to prevent adhesive from flowing into and clogging the microchannels. The second O-ring embedded in the interconnect ports is used to provide a tight seal when external capillaries are inserted into the interconnect ports. The pullout pressures of the capillary tubing depend on the diameter of the tubing and the number of O-rings. It was shown that with two Orings, the pullout pressure is about 2 MPa for tubing with inner diameters from 0.25 to 1 mm. Han et al. applied the SLA enabled interfacing technique in a multi-chip genetic sample preparation system [8]. The system consists of several functional components for cell purification, cell separation, cell lysis, solid phase DNA extraction, PCR, and capillary electrophoresis.
100
Wang and Shaikh
Fig. 3.5 Cross-sectional view of microfluidic interface fabricated using SLA by Han et al. [8] (This figure is reproduced with permission from Elsevier, ©2007, Elsevier).
Gonzalez et al. designed an on-chip fluidic tubing adaptor to interface their microfluidic system with the macro world [10]. The tubing adaptor is essentially a micromachined Si tube, which can connect to a microfluidic channel on one end and external plastic tubing on the other end, as shown in Figure 3.6. The authors fabricated the Si tube on the same chip as the microfluidic channels. Each hexagonal shaped Si channel consists of two halves fabricated simultaneously using anisotropic etching of a (100) silicon wafer on both sides. The two halves are joined together using fusion bonding methods. Plastic tubing with appropriate diameter can then be fitted over the hexagonal tube adaptor. Heat can then be applied at the junction to create a leak-free seal. To provide structural support and mechanical integrity, guide bars can be designed and fabricated on either side of the Si tube adaptor.
Fig. 3.6 Si based tubing interconnect developed by Gonzalez et al. [10] (This figure is reproduced with permission from Elsevier, ©1998, Elsevier).
Interfacing Microfluidic Devices with the Macro World
101
Gray et al. used the deep reactive etching method to create fluidic couplers in a silicon substrate for interfacing microfluidic systems with the macroscopic world [11,12]. The accuracy of microfabrication techniques allows high-density fluidic interconnections of arbitrary shapes and sizes to be made with precision. The authors reported three types of fluidic couplers made using this method. A first type of coupler was made by etching circular recess holes into the silicon substrate. The recess holes are connected to the micro channels in a microfluidic system. Since the size of the recess holes matches the diameter of capillary tubing, standard capillaries can be inserted into the recess holes and fixed in place using epoxy adhesives (Figure 3.7a left). Fluids can be transferred into the microfluidic systems through the capillaries using syringe pumps. To avoid possible seeping of the adhesive into the fluidic channel, a second type of fluidic coupler with a circular sleeve structure inside the recess hole was designed. The gap between the outer sleeve wall and the sidewall of the recess hole matches the thickness of the capillary tubing (Figure 3.7a right). When the capillary is plugged into the coupler, the sleeve serves as a barrier to the adhesive and enhances the mechanical robustness of the interface, though introducing some dead volume.
Fig. 3.7 Interconnect coupler structures developed by Gray et al. [11]. (a) Couplers requiring the use of adhesive. Capillary is placed directly in an etched hole (left) or in an etched circular sleeve (right). (b) Coupler using a plastic fitting to hold capillary in place (These figures are reproduced with permission from Elsevier, ©1999, Elsevier).
To eventually circumvent the use of adhesives, a third type of fluidic coupler was developed utilizing plastic press fittings, shown in Figure 3.7b. In this configuration, the capillary tubing is held in place by the plastic fittings. The plastic fitting can be made using injection molding of polyoxymethylene (POM). To integrate the plastic fittings into the system, additional through wafer holes are made using deep reactive ion etching or ultrasonic drilling methods. When the plastic fittings are assembled with the microfluidic chip, plastic pegs are melted with a heat gun to achieve
102
Wang and Shaikh
permanent assembly. The authors demonstrated the functionality of these microfluidic interfacing methods in a multi-level laminating mixer system. Pan et al. developed an adhesive-free method to form interconnections between microfluidic systems and the external world [13], as shown in Figure 3.8. In this method, a silicon flange is first created, which consists of s through wafer channel surrounded by a concentric sleeve structure. The silicon flange is formed by a two-step etching process including an anisotropic DRIE and an isotropic silicon reactive ion etching (RIE). Only one mask is needed to form the silicon flange and the RIE lag effect is used to differentiate the etch rate at the through wafer channel and the flange sleeve region. To interface with the external world, a silastic tube is connected to the silicon flange using a polyolefin heat shrink tubing. The polyolefin tube has a shrink temperature of 143oC and an expansion ratio of 2:1. When heated above the shrink temperature, the inner diameter of the polyolefin tube shrinks to half of its original size, forming tight seals at both the silicon flange and silastic tube ends. Pullout tests showed that the adhesion between the heat shrink tube and the silicon flange was dependent on the shape of the circular silicon sleeve structure. Among different shapes, circular sleeves showed maximal pullout forces of ~3N, due to large contact area between silicon sleeve and polyolefin tube. Leakage test showed that the heat shrink tubing connection had zero leakage at up to 29 psi for over 24 hours. Moreover, the polyolefin tubing becomes soft when heated to above 180oC and can be removed from the silicon flange, making the silicon microfluidic chip reusable. Since this method does not use any liquid adhesives, clogging of microfluidic channel due to adhesive seepage can be avoided.
Fig. 3.8 Schematic of adhesive-free microfluidic interconnect developed by Pan et al. [13] (This figure is reproduced by permission of the IEEE, ©2006 IEEE).
Interfacing Microfluidic Devices with the Macro World
103
3.3.2 Chip-to-world interfaces In many cases microfluidic technology can be useful for simplifying or optimizing a task that involves delivery of fluids to a macro-sized object. Whether this object is an analytical instrument or the human body, microscale systems offer advantages in terms of automation, fluid volume, and added functionality. Efficient connections between the chip and macro world are vital to the success of these microsystems. This section explores two major examples of chip-to-world interfacing: chip-based electrospray for mass spectrometry, and drug delivery/injection. 3.3.2.1 Mass spectrometer interfacing
A very important analytical tool in the life sciences is the mass spectrometer (MS). It can provide accurate determination of a molecule’s molecular weight, requiring only a few thousand molecules. A key parameter in the sensitivity of MS is the efficiency of ionizing the molecules in a sample. Conventional electrospray ionization techniques often form large fluid droplets that require high voltages in order to ionize completely. A major motivation for moving towards the micro-scale is the reduced volume and ability to integrate other pre-MS functionalities (e.g., liquid chromatography or capillary electrophoresis). A smaller sample volume (smaller droplet) allows for a smaller ionization voltage, which then allows the chip’s emitter to be placed closer to the MS ionization inlet. Also, the sensitivity may be increased due to the greater efficiency of ionizing a smaller sample volume. Chip-based MS interface development can be classified into three main areas: (1) electrospray directly from the chip edge, (2) mated capillary emitter, and (3) integrated emitter formed during the microfabrication process [14]. Initial attempts at interfacing microfluidic channels with MS utilized chip edge emitters, since fabrication was fairly straightforward. Karger’s group at Northeastern University was the first to publish results with this concept [15]. They created channels in glass chips via standard photolithographic and wet etching procedures. A total of nine individual channels were built with dimensions of 60-µm wide by 25-µm deep. Each channel extended right to the edge of the glass chip (Figure 3.9). Applying a high voltage between the sample inlet of one channel and the ionization inlet of the MS caused electrospray from the channel outlet. Due to buffer conditions (low electrical resistance or low pH), electroosmotic flow was small and they found it necessary to augment the electrical injection with pressure driven flow. A major issue with the chip edge emitter design is the spreading of
104
Wang and Shaikh
fluid along the edge of the chip, which is dependent upon the polarity of the fluid and hydrophillicity/hydrophobicity of the edge. This leakage could contaminate neighboring channels and effectively increases the fluid volume at the emitter, counteracting the benefits of moving to the microfluidic scale. If capillary electrophoresis (CE) was integrated on the device, the increased volume at the edge would recombine the separated fractions and render the CE useless. Xue et al. minimized the spreading issue for aqueous solutions by coating the chip edge with a reagent (i.e. Imunopen or n-octyltriacetoxysilane) that made the surface hydrophobic.
Fig. 3.9 Diagram of a system for performing chip edge electrospray. (This figure is reprinted with permission from [15]. Copyright 1997 American Chemical Socety).
Ramsey et al. implemented a chip edge emitter design and additionally integrated an electroosmotic pumping scheme [16]. As shown in Figure 3.10, a side-arm channel was placed very close to the edge emitter and was coated with a linear acrylamide to prevent electroosmotic flow within the channel (increased surface viscosity). By applying a 1.2-kV potential difference between the sample inlet and side-arm channel, electroosmotic flow was induced. The MS ionization inlet was grounded resulting in electrospray from the edge emitter due to a 4.8-kV potential difference between the side-arm channel and MS inlet. Separation between the chip edge and MS inlet was between 3-mm to 5-mm. The chip was scored and physically cleaved to produce the edge emitter. This method could possibly leave a rough edge and lead to locally high electric fields, which would yield a non-uniform electrospray.
Interfacing Microfluidic Devices with the Macro World
105
Fig. 3.10 (a) Schematic of a chip edge electrospray emitter with a side-arm channel to facilitate electrosmotic pumping. (b) Photograph of electrospray in action. (These figures are reprinted with permission from [16]. Copyright 1997 American Chemical Society).
Fig. 3.11 PET device for chip edge electrospray. (These figures are reprinted with permission from [17]. Copyright 2001 American Chemical Society).
As shown in Figure 3.11, Rohner et al. used a laser ablation process to fabricate an edge-based electrospray emitter in plastic (PET) [17]. A carbon ink electrode was embedded in the structure for electrospray operation. Since PET is hydrophobic, droplet formation at the chip edge was minimized for aqueous solutions. The use of photoablation and plastic lamination fabrication methods allow for low cost production and one-time-use devices. Plastics are not as optimal for CE separations as glass, so surface modifications may be necessary in order to integrate CE.
106
Wang and Shaikh
Fig. 3.12 Chip edge electrospray device integrating a nebulizer and CE separation. (This figure is reprinted with permission from [18]. Copyright 1999 American Chemical Society).
Zhang et al. took the initial work from Karger’s group a step further by integrating a nebulizer to minimize the droplet formation at the glass chip’s edge emitter outlet [18]. As shown in Figure 3.12, two auxiliary channels met at the exit of the main channel, constricting the droplet via pressurized flow from those channels. These channels also allowed for buffer modification prior to electrospray. They demonstrated on-chip CE separation and optimized electrospray emission into the MS inlet. By implementing a nebulizer and eliminating droplet formation on the chip edge, CE was achievable. Band broadening caused by serpentine channels and nonoptimal auxiliary channel design resulted in MS efficiencies lower than conventional methods. Design optimizations would likely increase the efficiency. In order to avoid the issues encountered with chip edge emitters, researchers have invented ways to couple glass capillaries to the microfluidic devices. Harrison’s group at University of Alberta drilled into the side of their chip to form a 200-µm diameter cavity into which a fused silica capillary was inserted [5]. They found that drill bits with flat ends could create holes optimal for capillary mating with low dead volume. Using this method they were able to integrate on-chip capillary electrophoresis with subsequent electrospray through the attached capillary and MS analysis (Figure 3.13) [19]. The manual drilling and assembly associated with this method are inherently error-prone and not optimal for mass production, which limits its use for industrial applications.
Interfacing Microfluidic Devices with the Macro World
107
Fig. 3.13 CE device mated with a glass capillary for electrospray. (This figure is reprinted with permission from [5]. Copyright 1999 American Chemical Society).
Recent work has focused on developing emitters that are formed during the microfabrication process and overcome the issues of chip edge emitters. Tang et al. demonstrated laser machining in polycarbonate to form a 2D array of nine emitters (Figure 3.14) [20]. These cone-shaped structures avoid large droplet formation and the hydrophobicity of the plastic further prevents liquid spreading. Emission results indicated an increase in MS sensitivity due to the array format since more ions were generated per run.
Fig. 3.14 2D array of cone-shaped emitters. (This figure is reprinted with permission from [20]. Copyright 2001 American Chemical Society).
As shown in Figure 3.15, Kameoka et al. sandwiched a triangular parylene film in between two laminated Zeonor sheets (a polyolefin) [21]. This triangular structure acted as a wick to direct fluid flowing out of microfluidic channels in the Zeonor film. A stable taylor cone was achievable with this device. The Zeonor channels were formed by hot embossing. Manual alignment was necessary to position the parylene tip prior to bonding the
108
Wang and Shaikh
two Zeonor sheets. Kameoka also demonstrated an array of tips, but each tip had to be separately handled and aligned. The manual nature of the process limits its use as a commercial solution.
Fig. 3.15 Triangular tip emitter sandwiched between two Zeonor sheets containing microfluidic channels. (These figures are reprinted with permission from [21]. Copyright 2002 American Chemical Society).
Yin et al. developed a polyimide-based device integrating a liquid chromatography column, enrichment column, and nanoelectrospray emitter tip [22]. The device was fabricated using laser ablation and is being sold by Agilent for use in its HPLC LC/MS instrument (Figure 3.16). Both the use of polyimide (plastic) and the triangular tip produce optimal conditions for the electrospray operation. The addition of sample enrichment (concentration) and on-chip chromatographic separation greatly facilitates the ability to rapidly perform MS analysis on complex samples.
Fig. 3.16 Polyimide device integrating chromatography, sample enrichment, and electrospray. (This figure is reprinted with permission from [22]. Copyright 2005 American Chemical Society).
Interfacing Microfluidic Devices with the Macro World
109
A few groups have focused on developing specialized nozzle emitter tips (Figure 3.17) [23-26]. Of these, the work from Shultz et al. has been successfully commercialized by Advion BioSciences [26]. The nozzle is fabricated in silicon using DRIE and can be as small as 15-µm in diameter. A sample delivery tube interfaces with the back of the chip, transporting the sample directly to the tip. Nozzles allow for minimal droplet formation and can accommodate liquids with varying polarity. Advion developed an array of nozzles and an automated instrument to selectively electrospray through one tip at a time. This design allows each tip to be one-time-use, provides a means to analyze multiple samples using the same chip, and results in a cost-effective, mass-producible commercial product.
a)
b)
c)
d)
Fig. 3.17 Nozzle emitter tips. (a) [23] (This figure is reproduced by permission of the IOP), (b) [24] (This figure is reproduced by permission of The Royal Society of Chemistry), (c) (This figure is reprinted in part with permission from [25]. Copyright 2003 American Chemical Society), (d) (This figure is reprinted in part with permission from [26]. Copyright 2000 American Chemical Society).
3.3.2.2 Chip-based drug delivery/injection
Microfluidics and microelectromechanical systems (MEMS) are attractive options for drug delivery systems since they are capable of providing precise dosage and spatial control. Much research has focused on developing microneedles for extracting fluids (i.e. blood) from the human body. Compared to conventional needles, microneedles produce much less pain in the patient due to their small size. Delivering fluids using microneedles or other methods has begun to gain interest as well. The major risk associated
110
Wang and Shaikh
with microneedles is the possibility of breaking the needle during insertion or removal (leaving the needle embedded in the skin). By designing the needle carefully, the risk can be minimized or eliminated. This section presents a few examples of microfluidic coupling to biological tissues for delivery of fluids. Paik et al. created in-plane silicon microneedles integrated with a PDMS microfluidic chip (Figure 3.18) [28]. Fabrication consisted of etching channels in the silicon via DRIE, sealing them by depositing a thick polysilicon layer, and then etching again to form the needle structure. Testing different needle geometries revealed that a 30 degree taper and isosceles triangle shape resulted in optimal resistance to mechanical failure under buckling load and needle insertion. A linear array of needles was shown to be effective at delivering a dye into model systems.
Fig. 3.18 Top: Array of silicon microneedles with buried microchannels. Bottom: The array was used to demonstrate successful delivery of Rhodamine B into a model system [28] (This figure is reproduced with permission from Elsevier, ©2004, Elsevier).
Interfacing Microfluidic Devices with the Macro World
111
Fig. 3.19 Silicon microneedle with integrated fluid conduits formed in photosilicate glass [29] (This figure is reproduced by permission of The Royal Society of Chemistry).
Zappe et al. fabricated a silicon microneedle attached to a pyrex wafer with etched microfluidic channels, as shown in Figure 3.19 [29]. Conduits 6.1µm and 2.3µm tall inside the needle were formed with a sacrificial photosilicate glass (PSG) layer. The needle was combined with an automated system for injecting drosophila embryos with RNA interference (RNAi) probes, resulting in a 10-fold efficiency increase compared to manual methods.
Fig. 3.20 Parylene microneedle with gold electrodes for fluid stimulus delivery and neural recording [30] (This figure is reproduced by permission of the IEEE, ©2006 IEEE).
Ziegler et al. used molding and thermal bonding of parylene to create a microfluidic channel integrated with a flexible neural probe [30]. As shown in Figure 3.20, integrated gold electrodes provide neuron recording capabilities for the envisioned drug response analysis system. Fluids could be delivered through the needle to neurons in the brain while measuring neuronal response.
112
Wang and Shaikh
Blake et al. built a PDMS perfusion chip to accurately deliver stimuli to specific regions of a neonatal rat brain slice (Figure 3.21) [31]. Conventional perfusion systems lack the ability to controllably vary the stimuli across the tissue. By immobilizing the brain slice on microfabricated PDMS posts and flowing solutions through the PDMS channels, the stimuli could be spatially controlled. Electrodes attached to the tissue allowed for real-time monitoring of the response to the stimuli.
Fig. 3.21 PDMS device for perfusion of an immobilized brain slice and stimuli response monitoring [31] (This figure is reproduced by permission of The Royal Society of Chemistry).
3.4. Future Perspectives As indicated by the various examples presented in this chapter, microfluidic interfacing has come a long way. The next step is to move from research level interfaces to more robust, easily manufacturable, and simple to use connections. While there are a few successful examples in industry,
Interfacing Microfluidic Devices with the Macro World
113
microfluidic interfacing is often a major hurdle that is not easily overcome. In order for a microfluidic device to move towards commercialization, this perennial problem must be solved. Commercial devices often need to be disposable such that the chip must have the capability to be inserted and removed from an instrument regularly. The interface between the chip and instrument must therefore be straightforward to operate while maintaining a robust connection.
References 1. Fredrickson CK, Fan ZH (2004) Macro-to-micro interfaces for microfluidic devices. Lab on a Chip 4:526-533. 2. Delamarche E, Juncker D, Schmid H (2005) Microfluidics for processing surfaces and miniaturizing biological assays. Advanced Materials 17:2911-2933. 3. Nittis V, Fortt R, Legge CH, de Mello AJ (2001) A high-pressure interconnect for chemical microsystem applications. Lab on a Chip 1:148-152. 4. Oh KW, Park CS, Namkoong K, Kim J, Ock KS, et al. (2005) World-to-chip microfluidic interface with built-in valves for multichamber chip-based PCR assays. Lab on a Chip 5:845-850. 5. Bings NH, Wang C, Skinner CD, Colyer CL, Thibault P, et al. (1999) Microfluidic devises connected to fused-silica capillaries with minimal dead volume. Analytical Chemistry 71:3292-3296. 6. Li SF, Chen SC (2003) Polydimethylsioxane fluidic interconnects for microfluidic systems. IEEE Transactions on Advanced Packaging 26:242-247. 7. Chiou CH, Lee GB (2004) Minimal dead-volume connectors for microfluidics using PDMS casting techniques. Journal of Micromechanics and Microengineering 14:1484-1490. 8. Han KH, Frazier AB (2005) Reliability aspects of packaging and integration technology for microfluidic systems. IEEE Transactions on Device and Materials Reliability 5:452-457. 9. Han KH, McConnell RD, Easley CJ, Bienvenue JM, Ferrance JP, et al. (2007) An active microfluidic system packaging technology. Sensors and Actuators B-Chemical 122:337-346. 10. Gonzalez C, Collins SD, Smith RL (1998) Fluidic interconnects for modular assembly of chemical microsystems. Sensors and Actuators B-Chemical 49:40-45. 11. Gray BL, Jaeggi D, Mourlas NJ, van Drieenhuizen BP, Williams KR, et al. (1999) Novel interconnection technologies for integrated microfluidic systems. Sensors and Actuators a-Physical 77:57-65. 12. Gray BL, Collins SD, Smith RL (2004) Interlocking mechanical and fluidic interconnections for microfluidic circuit boards. Sensors and Actuators aPhysical 112:18-24.
114
Wang and Shaikh
13. Pan T, Baldi A, Ziaie B (2006) A reworkable adhesive-free interconnection technology for microfluidic systems. Journal of Microelectromechanical Systems 15:267-272. 14. Koster S, Verpoorte E (2007) A decade of microfluidic analysis coupled with electrospray mass spectrometry: An overview. Lab on a Chip 7:1394-1412. 15. Xue QF, Foret F, Dunayevskiy YM, Zavracky PM, McGruer NE, et al. (1997) Multichannel microchip electrospray mass spectrometry. Analytical Chemistry 69:426-430. 16. Ramsey RS, Ramsey JM (1997) Generating electrospray from microchip devices using electroosmotic pumping. Analytical Chemistry 69:1174-1178. 17. Rohner TC, Rossier JS, Girault HH (2001) Polymer microspray with an integrated thick-film microelectrode. Analytical Chemistry 73:5353-5357. 18. Zhang B, Liu H, Karger BL, Foret F (1999) Microfabricated devices for capillary electrophoresis-electrospray mass spectrometry. Analytical Chemistry 71:3258-3264. 19. Li JJ, Thibault P, Bings NH, Skinner CD, Wang C, et al. (1999) Integration of microfabricated devices to capillary electrophoresis-electrospray mass spectrometry using a low dead volume connection: Application to rapid analyses of proteolytic digests. Analytical Chemistry 71:3036-3045. 20. Tang KQ, Lin YH, Matson DW, Kim T, Smith RD (2001) Generation of multiple electrosprays using microfabricated emitter arrays for improved mass spectrometric sensitivity. Analytical Chemistry 73:1658-1663. 21. Kameoka J, Orth R, Ilic B, Czaplewski D, Wachs T, et al. (2002) An electrospray ionization source for integration with microfluidics. Analytical Chemistry 74:5897-5901. 22. Yin NF, Killeen K, Brennen R, Sobek D, Werlich M, et al. (2005) Microfluidic chip for peptide analysis with an integrated HPLC column, sample enrichment column, and nanoelectrospray tip. Analytical Chemistry 77:527-533. 23. Griss P, Melin J, Sjodahl J, Roeraade J, Stemme G (2002) Development of micromachined hollow tips for protein analysis based on nanoelectrospray ionization mass spectrometry. Journal of Micromechanics and Microengineering 12:682-687. 24. Schilling M, Nigge W, Rudzinski A, Neyer A, Hergenroder R (2004) A new on-chip ESI nozzle for coupling of MS with microfluidic devices. Lab on a Chip 4:220-224. 25. Svedberg M, Pettersson A, Nilsson S, Bergquist J, Nyholm L, et al. (2003) Sheathless electrospray from polymer microchips. Analytical Chemistry 75:3934-3940. 26. Schultz GA, Corso TN, Prosser SJ, Zhang S (2000) A fully integrated monolithic microchip electrospray device for mass spectrometry. Analytical Chemistry 72:4058-4063. 27. NanoMate Fact Sheet. Advion BioSciences. (2008). 28. Paik SJ, Byun A, Lim JM, Park Y, Lee A, et al. (2004) In-plane single-crystalsilicon microneedles for minimally invasive microfluid systems. Sensors and Actuators a-Physical 114:276-284.
Interfacing Microfluidic Devices with the Macro World
115
29. Zappe S, Fish M, Scott MP, Solgaard O (2006) Automated MEMS-based Drosophila embryo injection system for high-throughput RNAi screens. Lab on a Chip 6:1012-1019. 30. Ziegler D, Suzuki T, Takeuchi S (2006) Fabrication of flexible neural probes with built-in microfluidic channels by thermal bonding of Parylene. Journal of Microelectromechanical Systems 15:1477-1482. 31. Blake AJ, Pearce TM, Rao NS, Johnson SM, Williams JC (2007) Multilayer PDMS microfluidic chamber for controlling brain slice microenvironment. Lab on a Chip 7:842-849.
Chapter 4 Genetic Analysis in Miniaturized Electrophoresis Systems
Li Zhu1 and Hong Wang2 1
GE Global Research, 1 Research Circle, Niskayuna NY 12309
2
Department of Chemistry, Louisiana State University, Baton Rouge, LA 70803
Correspondence should be addressed to: Li Zhu (
[email protected]) Hong Wang (
[email protected])
Keywords: microfluidics, microchip, electrophoresis, miniaturized electrophoresis system, genetic analysis, DNA analysis
Abstract As electrophoretic separation of DNA is one of the most important steps in many genetic analyses, continual advances in the development of miniaturized electrophoresis systems are critical to meet the growing need for highquality and low-cost genetic analyses. The commercialization of these miniaturized systems is having or will have significant impact on the revolutionary research in biomedical and life sciences areas. This chapter focuses on the microchip electrophoresis and its applications in genetic
118 Zhu and Wang
analyses. The recent developments in the fabrication of electrophoresis microchips, the sieving matrices, the novel micro-fabricated structures for DNA separation, and the internal wall coating chemistries are reviewed. A survey of recent progress in multiplexing and integrating steps toward a high-throughput, low-cost, and miniaturized electrophoresis system is provided. Some commercial microfluidic instruments for genetic analysis as well as their technologies are discussed and we conclude the chapter with a future perspective of the microfluidic technology in industry.
4.1 Introduction 4.1.1 Status of genetic analyses The Human Genome Project, with the primary goals of identifying all of the approximate 30,000-40,000 genes and determining the primary structure of the entire human genome comprised of its 3 billion base pairs, was initiated in 1990 and completed in 2003, the 50th anniversary of the discovery of the DNA double-helix structure [1, 2]. The successful completion of the Human Genome Project marked a significant milestone in the history of science. However, it does not signify an end to the further pursuit of novel sequencing and genotyping technologies. To the contrary, the demand for genomic sequence information, both in basic biomedical research and in routine clinical healthcare, has never been greater. This has stimulated a new round of competition in innovating high-throughput and cost-effective sequencing and genotyping strategies. In 2005, the National Human Genome Research Institute sponsored several research projects with the near term goal of cutting the sequencing cost of a mammalian sized genome to $100,000, and the final goal of a $1000 genome. DNA sequencing in a microchip by the Sanger method is one of the most promising approaches to achieve the near-term goal [3-6]. Currently, the cost of sequencing a mammalian sized genome with the state-of-the-art capillary array sequencer is around $1 million, which is still cost-prohibitive for most institutions and researchers. As alternatives, a variety of genetic analysis technologies targeting the variations in the genome have been explored. Instead of sequencing every single base along the genome, these technologies only interrogate the sites of interest, which have already been identified as the biomarkers for a specific disease or a specific application. DNA microarrays and electrophoretic separations are
Genetic Analysis in Miniaturized Electrophoresis Systems 119
two complementary methods and the most important approaches in genetic analysis. DNA microarrays, first developed in mid-1990s [7-12], turned out to be a powerful tool for high throughput genetic analysis. They interrogate hundreds to thousands of variations among a number of samples simultaneously. Microarray technology, however, will not be covered in the scope of this chapter. A few recent reviews on microarray technology can be recommended to readers for a quick glimpse of the state-of-the-art of the technology [13-21]. Complementary to microarrays, the size and/or conformation based electrophoretic separations of DNA generally target only one or a few variations in several to hundreds of samples. They are widely used in clinical diagnostics, forensic applications and biomedical research [22-29]. The last two decades have witnessed the transformation of the electrophoresis platform from the slab-gel to the capillaries/capillary arrays and to the present microchip format. This chapter will focus on the recent technology innovations in microchip/microfluidic electrophoresis for genetic analysis. 4.1.2 Genetic analysis by miniaturized electrophoresis system 4.1.2.1 Sanger sequencing of DNA
The Sanger sequencing method has served as the cornerstone for most genome sequencing efforts since first demonstrated by Frederic Sanger in 1977 [30]. It is a polymerase enzyme based replication method, similar to that in PCR, but using some specially designed 2’, 3’-dideoxynucleotides (ddNTPs) instead of all deoxyribonucleosides (dNTPs). The ddNTP terminate the extension reaction at a specific base when they are incorporated into the DNA chain due to the lack of an OH group on the 3’ –carbon position of the deoxyribose sugar. The termination occurs randomly at different positions in different copies, producing a nested set of DNA fragments of different length that are complementary to the unknown sequence (Fig. 4.1). After removing the excess ddNTPs and buffer components in the extension mixture by a purification step, the DNA ladders are electrophoretically separated. By correlating the length of the fragments with the identity of the terminating base through a fluorescently labeled dye, one can determine the nucleotide sequence using electrophoresis. Although various emerging sequencing technologies have been explored for ultra-low cost sequencing, the Sanger sequencing method remains the only viable technology for large genome de novo sequencing [3-5]. Tremendous efforts have been devoted to reducing the reagent consumption
120 Zhu and Wang
and increasing the throughput by developing multiplexed and/or integrated electrophoresis microchips in Sanger sequencing. Recently, a read length from a glass electrophoresis microchip was reported comparable to that of conventional capillary arrays, while the throughput was several folds higher and the reagent consumption was one to two orders of magnitude less [31]. Disposable plastic microchips were also tested for Sanger sequencing. However, it is still challenging to achieve the required single base resolution and long read length on those microchips [32, 33].
Fig. 4.1 Top: Chemical strudctures of deoxyribnucleoside (dNTP) triphosphate and dideoxyribonucleoside (ddNTP) triphosphate. Bottom: Schematic diagram of Sanger chain termination DNA sequencing reaction.
Genetic Analysis in Miniaturized Electrophoresis Systems 121 4.1.2.2 Genotyping: polymorphisms and mutations
More than 99% of human DNA sequences are identical across the population. However, the genetic or DNA sequence variations in the genome can have a major impact on predisposition to diseases, response to environmental factors, and the effectiveness of the medicines and vaccines [3436]. When the variations in the genome are present in at least 1% of the population, they are generally not harmful and are called polymorphisms. Otherwise, when the variations are less common (less than 1%) and frequently result in diseases, they are called mutations. A variety of size and/or conformation based electrophoretic separation strategies have been developed for identifying polymorphisms and mutations in the human genome during the past decades [28, 37]. Single nucleotide polymorphism (SNP), a single base substitution, deletion or insertion, is the most abundant variation in the human genome. There are more than 3 million SNPs that have been identified, which corresponds to about 1 in every 1000 bases. Several electrophoretic separation strategies have been used to identify the SNPs and single base mutations, including: restriction fragment length polymorphism (RFLP), single strand conformational polymorphism (SSCP), and heteroduplex analysis (HA). RFLPs are applied when the genetic variations create or delete the sites that can be recognized by specific restriction enzymes. The difference in homologous DNA sequences can be detected by the presence of the differently sized fragments after the restriction endonuclease digestion and the electrophoretic separation. RFLPs can be used in paternity testing and disease diagnosis. Guttman et al. developed a rapid PCR-RFLP method to analyze the mitochondrial DNA mutation in diabetes [38]. Microchip electrophoresis was used by Qin et al. to genotype the -6A/G polymorphism in the core promoter region of the AGT gene using the PCR-RFLP method [39]. SSCP is a simple and versatile method of electrophoretically separating the single-stranded DNA (ssDNA) based on subtle differences in sequence (often a single base pair). While the mobility of double-stranded DNA (dsDNA) depends largely upon the fragment sizes but little on the fragment composition, the mobility of ssDNA can be dramatically influenced by the nucleotide composition of a fragment with a specific three dimensional conformation resulting from the intra-strand base pairing. Tian et al. used microchip electrophoresis to detect the common mutations in the BRCA1 and BRCA2 gene [40]. A 384-lane electrophoresis microchip was tested in genotyping 21 single-nucleotide polymorphisms (SNPs) from
122 Zhu and Wang
HFE, MYL2, MYL3, and MYH7 genes associated with hereditary hemochromatosis (HHC) and hereditary hypertrophic cardiomyopathy (HCM) [41]. The HA method is based on the mobility difference between the heteroduplex and the corresponding homoduplex. It is a simple and convenient method that has been widely used in biomedical research and clinical diagnostics [42]. In addition, HA can be combined with SSCP and allelespecific amplification method for rapid identification of mutations [43-46]. 4.1.2.3 Short tandem repeats analyses
Short tandem repeats (STRs) or microsatellites are another common polymorphism predominantly used in forensic identification and paternity testing [23, 47, 48]. STRs are 7 to 20 repeats of specific DNA sequences ranging from 2 to 10 bases long. Recently, a 96-channel microfabricated capillary array electrophoresis device was evaluated for forensic STR typing using commercially available forensic kits, e.g., PowerPlex 16 and AmpFlSTR Profiler [49]. In most cases, only 2-3 base-pair resolution is required for STR analysis, therefore plastic microchips are well-suited for such an application. Shi et al. reported a successful two-color sizing analysis of four-locus (CSF1PO, TPOX, TH01, vWA) STRs in poly (cyclo olefin) microchips [33, 50].
4.2 Microchip Electrophoresis for Genomic Analysis Microchip electrophoresis shares basic principles with slab-gel and capillary electrophoresis, therefore, the DNA migration models, the gel matrices, the surface coating strategies and the fluorescent dyes developed for slab-gel and capillary electrophoresis systems are suitable for microchip electrophoresis as well. Microchip electrophoresis also has its unique properties and advantages, such as its ultra-fast analysis time, ultra-low sample consumption, capability of easy scale-up for high throughput analysis and potential for convenient system automation and integration. More recently, various thermoplastic materials have been explored for the mass production of disposable electrophoresis microchips, making lowcost microchip-based instruments more practical.
Genetic Analysis in Miniaturized Electrophoresis Systems 123
4.2.1 Material and fabrication of electrophoresis microchips 4.2.1.1 Glass, quartz, and silicon
Glass, quartz, and silicon are the most widely used materials in microchip electrophoresis due to their high-quality optical properties, welldocumented surface chemistries and well-developed fabrication procedures adopted from the microelectronic industry [51-57]. These microchips are mainly manufactured by standard photolithographic procedures followed by chemical etching. Unfortunately, the fabrication of these glass or quartz-based micro-devices is relatively expensive and time-consuming; which make them less desirable for low cost applications. In addition, isotropicity of the wet etching in conventional fabrication produces shallow, elliptical shaped structures with low aspect ratios. Although high aspect ratio structure can be made by a dry etching method, such as DRIE (dry reactive ion etching), the high cost and long fabrication time are still costprohibitive. Other disadvantages associated with microchips made from these materials include the requirement of high temperatures (600 °C) and the use of harmful chemicals (e.g., hydrofluoric acid). 4.2.1.2 Polymer/plastic materials
To circumvent the limitations associated with glass-based devices, several polymer materials have been explored as the microchip substrates. Polymers offer a wide range of mechanical and thermal properties, providing various selections for different applications and allowing rapid inexpensive mass production of micro-devices. They make the economical single-use devices become possible, eliminating cleaning and sample-to-sample carryover contamination. PDMS (polydimethylsiloxane) is the most popular elastomer used for microchip fabrication [58, 59]. It is an inexpensive, flexible and chemically inert material with optically transparency down to 230 nm and little autofluorescence. The PDMS microchannels can be fabricated by softlithography [59]. Briefly, a negative mold is fabricated by conventional micromachining technology. For example, the negative thick photoresist SU-8 has been widely used to fabricate negative mold through photolithography. Then the PDMS prepolymer (e.g., the 10:1 mixture of base and curing agent) is casted over the mold. After curing, which typically takes a few hours, the PDMS film carrying the microstructure can be peeled off from the mold. This PDMS film can be reversibly sealed to another PDMS film or other ultra-flat surfaces by conformal contact. It can also be irre-
124 Zhu and Wang
versibly sealed to another PDMS film or a glass plate after an oxygen plasma treatment. Because of the elastomeric nature, PDMS deforms easily. An interesting and effective approach to overcome this is to fabricate PDMS-glass or PDMS-silicon hybrid microchips, which not only takes advantage of the easy and fast fabrication of PDMS, but also the strength from the rigid glass or silicon materials [60-62]. Thermoplastics are attractive alternatives for mass production of the low cost and disposable microchips as well. The thermoplastics used in electrophoresis microchips include PMMA (polymethylmethacrylate), COC (cyclic olefin copolymers, commercialized as Topas® by Ticona and Zeonex® by Zeon), and PC (polycarbonate). One of the most attractive features associated with thermoplastic materials is their flexibility in the micromachining process [63-67]. Thermoplastic fabrication techniques can be broadly classified into two categories, direct fabrication methods and replication methods. Direct methods, by which the individual polymer surface is fabricated to form microstructures, are mostly used for rapid prototyping. Examples include mechanical milling and laser ablation [68]. They are not the most economical and in some cases produce rough surfaces that are problematic for highresolution electrophoretic separations. Replication methods employ a precise template or molding tool from which identical features can be replicated. Examples are injection molding [69], hot-embossing [70], and imprinting [71]. Such techniques are proven to be cost-effective when large numbers of identical plastic parts or devices are formed. Hot-embossing is an easy and inexpensive method for reproduction of microstructures. Under vacuum conditions, a metal molding tool is pressed into a heated polymer substrate, transferring the desired features into the polymer. The number of daughter parts that can be fabricated from a single master and the average replication time per device is ~5 min. Injection molding is the industrial standard process for manufacturing macroscopic plastic parts, and it has been optimized for microstructure fabrication. The molten thermoplastic particles are transported to the heated mold cavity, which contains negative mold. Then the cavity will be cooled to release the microchip. By adjusting the process time and temperature, injection molding can be used to fabricate 3-D micro-structures with excellent precision. However, the high temperature and pressure during the fabrication process introduce thermal stresses, which make the microchip sensitive to the temperature and pressure fluctuations. Further material and fabrication methods for microchips can also be found in Chapter 2.
Genetic Analysis in Miniaturized Electrophoresis Systems 125
4.2.2 Theory of gel electrophoresis of DNA The electrophoretic mobility of DNA in free solution is independent of its size due to the constant charge-to-size ratio. In order to achieve a size separation of DNA, it is necessary to introduce certain physical obstacles on a molecular level. One of the widely used physical obstacles is called sieving matrix or gel, which will selectively retard the migration of DNA molecules under an electric field based on their sizes and conformations. Numerous polymers or gels have been utilized as the sieving matrices in capillary and microchip electrophoresis. Understanding the interaction between the polymer networks and DNA species during electrophoretic separation is critical in optimizing sieving matrices and improving performance of electrophoresis [72-75]. Two widely accepted models of DNA electrophoresis in a gel polymer are the Ogston model and the reptation model. The Ogston model [76] treats the polymer network as a molecular sieve. It assumes that the matrix consists of a long, inert and randomly distributed network of interconnected pores with a certain average pore size, through which the solute molecules migrate as a spherical coil. These unperturbed spherical migrating species diffuse laterally until they encounter pores large enough to accommodate passage. Small molecules migrate faster because they have accessibility to a larger fraction of the available pores. The Ogston model predicts a linear relationship between the log µ, where µ is the electrophoretic mobility, and the gel concentration for small DNA molecules and low electronic field. However, according to Ogston model, the mobility of migrating species will quickly approach zero when the molecular radius approaches the pore size of the sieving matrix. In other words, this model cannot explain the behavior of large, flexible molecules migrating through sieving media with mesh size significantly smaller than their size. This phenomenon can be described by the reptation model. The reptation theory [77] suggests that large DNA molecules, when they are too large to fit through a pore while maintaining a coiled conformation, would exhibit “snake-like” migratory behavior moving through the smaller pore network of the gel under the influence of an electric field. This model is based upon the assumption that the migrating molecule can deform and stretch, behaving as a free draining coil instead of an immutable sphere with fixed radius of gyration described by the Ogston model. The mobility of the analyte molecule by this process is inversely proportional to its molecular size. When the electric field strength increases, reptation turns into a biased reptation model [78]. Under this model, the electric field-induced
126 Zhu and Wang
orientation extends the stretching periods of DNA, causing their random walk to become strongly biased in the forward direction. The leading end of DNA becomes completely aligned with the field so that no further field effects exist. In a fully biased reptation regime, the mobility increases to saturation at which point there is no dependence of mobility on DNA length and this results in all large fragments migrating at the same rate. With the fast growth of capillary electrophoresis in 1990s, the aforementioned two models derived from slab-gel electrophoresis have become less convincing when it comes to explaining the high sieving power a linear polymer possesses. There have been inconsistencies between theory and experimental observations. Barron et al. [79] showed that the separation of 2.0-23.1 Kbp DNA fragments is possible in 0.0006% (w/w) uncrosslinked hydroxyethyl cellulose (HEC) polymers. To explain the DNA separation in ultra-diluted and unentangled polymer solution, a transient entanglement coupling mechanism was proposed, which postulated that the DNA fragments hook and then drag the uncharged polymer chains during the migration. The larger DNA molecules have higher chance of encountering and entangling polymer molecules, therefore, experience a greater reduction in mobility. This mechanism also applies to DNA separation in capillary electrophoresis using linear entangled polymer solutions at high concentration. 4.2.3 Gel matrices The DNA migration behavior and the separation efficiency are largely affected by the sieving matrices. Therefore, the choice of the sieving matrices is critical to the success of the separation of DNA. Most of the gel matrices that are applied to capillaries can be applied to microchips as well, if they can be loaded into the microchannels. Compared with their conventional counterpart, microchips cannot sustain as much pressure as the capillary during the matrix loading, e.g. 50 psi for most plastic microchips and 200 psi for glass or silicon microchips. An ideal sieving matrix for DNA separation on a microchip should include the following properties: low viscosity, high molecular sieving power, dynamic coating ability, and cost effectiveness. Recent efforts to improve microchip DNA separations have led to the development of new sieving solutions that are more compatible with the microlfuidic format [80-83]. One example is the development of the thermoresponsive polymer [81]. It possesses lower viscosity for easy loading, and its viscosity changes with temperature to be suitable for separation. However, the optimal performance of such a polymer has not been
Genetic Analysis in Miniaturized Electrophoresis Systems 127
fully developed yet. The effort in designing and synthesizing an ideal polymer for DNA separation on a microchip is still continuing. In this section, the traditional gel matrices as well as some novel polymer matrices will be introduced. 4.2.3.1 Cross-linked polyacrylamide gel
Cross-linked polyacrylamide is one of the most commonly used separation matrices for slab gel DNA sequencing. It is prepared by a radical copolymerization of acrylamide, with a varying amount of N, N’-methylene bisacrylamide, a cross-linker. The polymerization is typically initiated by ammonium persulfate (APS) and catalyzed by N, N, N’, N’tetramethylethylenediamine (TEMED). Upon polymerization, a dense, cross-linked and flexible polymer network is formed, which has pore sizes ranging from a few nanometers to tens of nanometers. This tight pore network is effective for high-resolution separations of DNA ranging from 6 bases to about 1000 bases long, depending on the gel concentration and the level of cross-linking. Although having been long standing with the slab gel technology, cross-linked polyacrylamide is rarely used in capillaries or microchips. This is because of a few drawbacks associated with it, such as the gel shrinkage and the formation of bubbles during polymerization, as well as the degradation of the polyacrylamide by alkali hydrolysis during electrophoresis. 4.2.3.2 Polysaccharide and its derivatives
Polysaccharides and their derivatives have been used as sieving matrices for DNA separation in both capillary and microchip electrophoresis [83]. They are usually selected when the high-resolution separation is not required. Low viscosity cellulose derivatives, methylcellulose (MC), hydroxyethylcellulose (HEC), hydroxypropyl cellulose (HPC), and hydroxypropylmethylcellulose (HPMC) were used on PMMA microchips for ultra fast dsDNA separations. Agarose, the most widely used lowresolution slab-gel material, was used in a PDMS-PMMA hybrid microdevice, a glass-silicon, and a PDMS-glass hybrid microchip for double strand DNA separation [84-86]. 4.2.3.3 Linear polyacrylamide (LPA) and acrylamide derivatives
Since the first introduction of replaceable, ultra-high-molar-mass Linear Polyacrylamide (LPA) matrices by Karger et al. [87-90], LPA has been the matrix of choice for high performance DNA separation, and in particular,
128 Zhu and Wang
DNA sequencing. LPA gels can be formed by polymerizing acrylamide the same way as cross-linked polyacrylamide, except that no cross-linking agent is added. LPA possesses a few highly desirable properties for DNA separation, particularly its high hydrophilicity and excellent sieving power owing to the physical entanglements of the polymer chains. In addition, these linear polymers are not as rigid as their cross-linked counterpart. Therefore, they can be easily replaced in the capillary or microchip after each run. The first DNA sequencing by CE with replaceable linear polyacrylamide reported a read length of 350 bases from an M13 template in 30 minutes [88]. Later, Karger and co-workers reported an improved sequencing read length of 1300 bases in 2 hours [90] by using an LPA mixture composed of 0.5% w/w 270 KDa and 2% w/w 17 MDa based on their previous result of 1000 bases in less than 1 hour [87, 89]. Although numerous types of water soluble, high-molar-mass polymers have been investigated for sequencing applications, LPA remains the most widely accepted sieving matrix for high-performance DNA sequencing by both capillary and microchip electrophoresis. Despite the superior performance of LPA gels, their high viscosity and lack of intrinsic wall-coating ability make them less suitable for microchips. Their high viscosity makes them difficult to fill into and be removed from the microchannel, and in many cases, requires the application of high pressure. In addition, a rather complex coating procedure is usually required on the microchannel walls prior to electrophoresis to minimize the electroosmotic flow and the analytewall interactions. Barron and co-workers created the concept of “nanogel” by incorporating a low percentage of cross-linker in high-molar-mass LPA [91, 92]. The small amount of cross-linker localizes the cross-linking and keeps the flowing ability of LPA. It has been reported that the nanogel provides 18% longer read length in sequencing than a matched LPA counterpart in a microdevice. Poly-N, N-dimethylacrylamide (PDMA) is a widely used acrylamide derivative for DNA separations. This polymer is commercialized by Applied Biosystems as POP (Performance Optimized Polymer) gels. Advantages of PDMA over LPA are its resistance to hydrolysis, self-coating ability, and significantly lower viscosity. However, the performance of PDMA has been less satisfactory compared with that of LPA. This may in part due to the hydrophobic interaction between the PDMA and the DNA labeling dyes. The best performance achieved was 800 bases with a final resolution of 0.5 in 96 min, and 1000 bases with a final resolution of 0.3 in a capillary [93]. PDMA has been the matrix of choice for a few commercial micro-
Genetic Analysis in Miniaturized Electrophoresis Systems 129
chip electrophoresis instruments, such as the 2100 Bioanalyzer by Agilent (www.agilent.com), for which high-resolution DNA sizing (e.g., singlebase resolution) is not required. 4.2.3.4 Thermoresponsive polymer matrices
Thermoresponsive polymer matrices exhibit low viscosity at one temperature for gel loading and high viscosity at another temperature for high resolution separation [81, 82]. They can be divided into two categories: thermo-thickening (thermo associating) polymers and thermo-thinning polymers. Thermo-thickening or thermo associating polymers have low viscosity at the low temperature. The viscosity of the polymer increases when it is heated to the high critical solution temperature (HCST). Pluronic polyol F127 (PEO99PPO69PEO99) [94], a typical block copolymer with thermal self-association properties, was tested for dsDNA separation in a PMMA microchip. A grafted copolymer with a hydrophilic LPA backbone and comb-like pNIPA side chains, it was investigated for DNA sequencing. It delivered read length of 800 bases at the optimized composition [95]. Thermo-thinning polymers, on the other hand, have high viscosity at the room temperature. When heated up, the polymer becomes less viscous. Thermo-thinning random copolymers composed of different ratios of DEA (N,N-diethylacrylamide) and DMA (N,N-dimethylacrylamide) were used for DNA sequencing, in both capillary and microchip [81]. Thermoresponsive polymers are promising alternatives as sieving matrices for microchip electrophoresis; however, their performance still needs to be optimized. It would be difficult for these polymers to beat the superior performance of LPA, mainly due to the hydrophobic moieties required for the thermo-response. 4.2.3.5 Other Polymers: PEO, PVP, and PVA
Other replaceable polymer solutions, such as polyethylene oxide (PEO), polyvinyl pyrrolidone (PVP), and polyvinyl alcohol (PVA), have been developed and reported for both DNA fragment analysis and sequencing. PEO is a self-coating polymer widely used in DNA separation. It requires a surface pre-treatment as long as 2-hours with HCl to enhance the silica surface absorption before electrophoresis. Yeung’s group achieved a read length of up to 1000-base with PEO at the optimized conditions in a capillary [96, 97]. Xu et al. systematically investigated the performance of PEO
130 Zhu and Wang
in separating DNA ladders at various combinations of molecular weight and concentrations in PMMA microchips [98]. PVP is another polymer with self-coating ability but also a low viscosity. It can provide moderate separation efficiencies under optimum gel compositions. Several groups have demonstrated DNA fragment sizing in both capillaries and microchips using such a polymer [99-102]. The main obstacle for the routine use of PVP in DNA separation is the large mobility shifts of fluorescently labeled DNA fragments observed in this matrix. PVA is resistant to hydrolysis over all pH ranges, and it maintains very low viscosity even at concentrations up to 12%. It was reported as a sieving matrix for fragment analysis in a capillary [103]. However, its separation efficiency diminishes very quickly due to the strong hydrogen bonding presented in the polymer, resulting in self-aggregation. PVA can also be used as a coating polymer to reduce the EOF and analyte absorption onto the walls. 4.2.4 Novel DNA separation strategies on microchips Recently, alternative mechanisms for DNA separation without using a sieving matrix have been introduced for microchip electrophoresis. One approach is to label the DNA fragments with large uncharged molecules, e.g. streptavidin or long chain PEG, to break the constant charge to size ratio. The longer end-labeled DNA fragments migrate faster than the shorter ones, so the size dependent separation can be achieved in a free solution. This method is called end-labeled free-solution electrophoresis (ELFSE) [104, 105]. A number of novel separation strategies employing microfabricated mechanical obstacles at a size scale comparable to the radius of DNA gyration, including micro- and nano-pillar arrays, nano-spheres, and nano-sized channels, have also been developed. The idea of replacing the gel or polymer sieving matrices with micrometer or sub-micrometer sized pillar arrays was first demonstrated by Volkmuth and Austin, who fabricated a rectangular array composed of 1 µm diameter cylindrical posts with 2 µm center-to-center distance on a silicon wafer by optical microlithography [106, 107]. Comparable to the nominal pore size of a 0.05% agarose gel, this array could resolve the DNA fragments up to 100 Kbp. Similar arrays were also employed for pulsed-field electrophore-
Genetic Analysis in Miniaturized Electrophoresis Systems 131
sis [108] and continuous macromolecule sorting [109] by the same group. A much more complex and powerful demonstration of this array format was a “DNA prism” for high-speed continuous sorting of DNA fragments in different directions according to their sizes, just as an optical prism deflects the light at different angles according to the wavelength (Fig. 4.2) [110]. Under a symmetric pulsed field, the DNA prism could separate 61209 Kbp DNA molecules in 15 seconds with ~13% resolution, which is 1000 times faster than the conventional pulsed field gel electrophoresis (PFGE) (10-240 hours), and 40 time faster than the pulsed field capillary electrophoresis (~40 minutes). To improve the performance of small DNA fragment separation, Kaji et al. fabricated sub-micrometer, high aspect ratio nano-pillars (100-500 nm diameter and 500-500 nm tall) to resolve DNA fragments ranging from 1 to 38 Kbp within a 380 µm-long channel in 10 seconds [111].
Fig. 4.2 Structure of the microfabricated device illustrating the sieving matrix integrated with the microlfuidic channels. The post array is 3 mm by 9 mm, and the posts are 2 µm in diameter, 2 µm apart, and 2 µm tall. A single channel connecting to the DNA reservoir injects DNA through a 28-V opening. The many microfluidic channels connected to buffer reservoirs produce uniform electric fields over the sieving matrix by acting as electric-current injectors. (Reprinted by permission from Macmillan Publishers Ltd: Nature Biotechnology [110], copyright 2002.)
In contrast to the pillar arrays fabricated by conventional photolithography technology, Doyle et al. introduced a self-assembled, quasi-regular temporary pillar array by applying a constant and homogeneous magnetic field to a suspension of paramagnetic particles confined in a microfluidic channel. By alternating the size and the concentration of the magnetic particles, the
132 Zhu and Wang
nominal “pore” size of the array can be altered for different applications [112]. A separation of λ-phage, 2λ-DNA, and bacteriophage T4 DNA was achieved in as short as 150 s and resolutions greater than 2 between λ and T4 were successfully obtained [113]. In addition, the movement of a single T4 DNA molecule (169 Kbp) in the magnetic particle pillar array and a theoretical model were presented [114]. Brownian motion has been exploited by several groups for the size dependent DNA separation. One approach is to fabricate an asymmetric array of obstacles as a Brownian ratchet to rectify the Brownian motion laterally and thereby deflect diffusing particles depending on their size. In a 4-inch (10-cm) silicon wafer, a nominal 6% resolution by length of DNA molecules in the size range of 15–30 Kbp was demonstrated by Chou el al. in 1999 [115], followed by a continuous DNA sorting device by Cobadi et al. in 2002 [116]. To speed up the separation based on the intrinsically slow Brownian motion, a tilted electric field was applied to the array [117]. The time required for resolving 48.5 Kbp and 164 Kbp DNA fragments was reduced from 140 min to 14 min, a 10-fold improvement. Alternatively, a Brownian ratchet could be created by the electric field instead of physical obstacles, as demonstrated by Bader et al [118, 119]. More recently, another Brownian motion based phenomenon, absolution negative mobility (ANM), was explored for separating DNA fragments [120, 121]. Another intriguing approach is the entropic based separation, which was demonstrated by an “entropic trap” device [122] (Fig. 4.3) and an “entropic recoil” device [123, 124] (Fig. 4.4). The basic principle of entropic based separation is that DNA molecules are energetically favorable to be in the coiled state since this maximizes their entropy. The entropic trap device is a microfluidics channel defined by a series of alternating thin (75100 nm) and thick (1.5-3 µm) regions in the microchannel. The DNA molecule can keep in the coiled state in the thick region; however, when it meets the thin region where the channel size is much smaller than the gyration of a DNA molecule, the DNA has to deform to pass through. Since the deformation is entropic unfavorable, the DNA molecule is temporarily trapped at the entrance of the thin region. An escape of the whole DNA molecule is not initiated until majority of that molecule is introduced into the thin region by Brownian motion. Interestingly, the longer molecules have a larger contact area and consequently a higher probability to escape the entropic trap, therefore, they move faster than the shorter ones. The entropic recoil device consists of a high entropy pillar-free planar region and a low entropy pillar-filled region on a chip. After applied the electric field, the DNA molecules are forced to move to the pillar region. When the elec-
Genetic Analysis in Miniaturized Electrophoresis Systems 133
tric field is removed, the molecules will recoil to the planar region unless they have entirely moved to the pillar region. In contrast to the entropic trap, the shorter DNA molecules move faster than the longer ones. A rapid separation of T2 (167 Kbp) and T7 (39 Kbp) was demonstrated. Applying a voltage pulse of 2 s, the shorter T7 molecules were fully inserted, while the longer T2 molecules remained partially in pillar-free region and subsequently recoiled.
Fig. 4.3 Nanofluidic separation device with many entropic traps. (A) Crosssectional schematic diagram of the device. (B) Top view of the device in operation. (C) Experimental setup. (From [122]. Reprinted with permission from AAAS.)
134 Zhu and Wang
(a)
(b)
Fig. 4.4 (a) DNA molecules are driven into the dense-pillar region, which occupies the bottom 80% of each frame. (b) Entropic recoil at various stages ending when all the molecules have recoiled except the leftmost. (Reprinted figure with permission from [124]. Copyright (2002) by the American Physical Society.)
4.2.5 Surface coating methods for microchannel walls Similar to the conventional capillary electrophoresis, the surface properties of the internal walls of a microchip have a significant impact on the performance of the separation. Surface coating is usually required to reduce the electroosmotic flow (EOF) and to minimize the analyte-wall interactions, both of which deteriorate the separation efficiency by introducing band-broadening during gel electrophoresis. The nonuniform surface charge resulting from the material itself or/and the micromachining process, as well as the nonspecifically absorbed DNA molecules over the runs, affects the EOF during electrophoresis, and thus severely deteriorates the reproducibility. In addition, the large surface-to-volume ratio in a microchip makes it even more important to suppress the interaction between DNA and the internal wall, since the small volume of samples can easily be depleted, and hence detection sensitivity impaired. Surface coating strategies can be divided into two categories: dynamic coating (or physical adsorption) and permanent coating (or chemical modi-
Genetic Analysis in Miniaturized Electrophoresis Systems 135
fication) [125-129]. Compared to the often time-consuming and laborintensive chemical modification, dynamic coating is a simple, fast and low cost option, especially for disposable plastic microchips. But the permanent surface coating is usually more uniform and stable. Following the same categorization as microchip fabrications, both surface-coating chemistries will be reviewed in the following section based on different substrate materials. 4.2.5.1 Glass/silicon chip surface coatings
Both the dynamic and permanent (or physical and chemical) coating methods that have been well developed for fused-silica capillaries during the “capillary era” are fully transferable to glass/silicon microchips simply due to the fact that they are the same material and share the same surface chemistry. Various polymers, such as PDMA, PEO, PVP, PVA and HEC, have been identified not only as a sieving matrix, but also as a dynamic coating materials in both DNA fragment sizing and long-read DNA sequencing applications [125]. The “self-coating” is simply realized by physical adsorption and coating the microchip can be easily carried out by pumping the polymer matrices through the microchannel. Compared to the multi-step chemical modification procedures, dynamic coating is more amenable for automation and electrophoresis integration. On the other hand, the most common surface coating for glass microchips is still chemical modification because of its robustness and well-documented history. It involves the use of the classic Hiertén protocol [130] or its modified version, in which a neutral polymer layer is covalently bond to the glass/silicon surface activated by a standard silanization chemistry. For example, the protocol of covalently bonding the polyacrylamide through the γmethacryloxypropyltrimethoxysilane was optimized on the surfaces of glass microchips for both fragment analysis and sequencing applications [131, 132]. Chemically modifying the glass surfaces with a thin-layer of PVA coating was also demonstrated [133]. 4.2.5.2 Polymer/plastic chip surface coatings
PDMS is one of the most popular substrate materials for microfluidic chips. Unfortunately, the native PDMS is not ideal for electrophoresis due to its inherent strong surface hydrophobicity, because of which the microchannel is very difficult to wet and air bubbles are easily generated. Additionally, the strong tendency for analyte absorption produces unstable
136 Zhu and Wang
EOF. Numerous dynamic coating and chemical modification strategies have been proposed to tailor the surface properties of PDMS for different applications [127]. HPC and HPMC were used as the PDMS self-coating sieving matrices for dsDNA fragment analysis in a few early reports, though the authors did not clearly indicate the dynamic coating properties of the HPC and HPMC [134, 135]. PDMS surfaces can be chemically modified by silanization or UV grafting. When using silanization, PDMS surfaces are treated with oxygen plasma or Tesla coil to oxidize the Si-CH3 group to Si-OH group, followed by the similar silanization chemistry applied on glass/silicon surfaces [136, 137]. When using UV grafting, PDMS microchannels are filled with different monomers, including acrylic acid, acrylamide and PEG-monomethoxyl acrylate, etc. After exposure under a mercury lamp, these monomers are grafted to the PDMS surface to yield hydrophilic surfaces and produce stable EOF [138-140]. Surface modification on thermoplastics, including PMMA, PC, and COC has been substantially studied to make them ideal substitutes to glass microchips, offering not only the low cost, but also the high performance for electrophoresis [125]. However, each thermoplastic material has its unique functional groups at the surface. Unlike the universal silanization chemistry applied to glass and PDMS surfaces, a different chemistry has to be designed and optimized for each surface and those polymers may not be compatible with some of the organic solvents used during chemical modification. Under such circumstances, dynamic coating is especially attractive simply because the usually tedious chemical modification does not harmonize with the low cost and disposable nature of those chips. Several dynamic coating methods have been investigated for the surfaces of the thermoplastics. Hydrophilic neutral polymers including PEG, HEC, HPMC, and methylcellulose (MC) have been tested for PMMA surface dynamic coating. For example, HPMC and Pluronic polyol F127 (PEO99PPO69PEO99) were used as self-coating sieving matrices for dsDNA separation in PMMA microchip by Xu et al. [141] and Song et al. [94], respectively. DNA sequencing and STR genotyping were successfully demonstrated in Zeonor (poly(cycloolefin)) microchip dynamically coated with 2% poly(dimethylacrylamide/diethylacrylamide) [33].
Genetic Analysis in Miniaturized Electrophoresis Systems 137
Although not aligned with the low cost nature of the disposable plastic microchips, to demonstrate the feasibility of using those thermoplastic materials for electrophoresis, many chemically modification strategies have been pursued [142, 143]. The chemical modifications allow some improvement of the analytical performance of the plastic microchips but they are still not good enough for the high-resolution applications like DNA sequencing.
4.3 Parallelization in Microchip Electrophoresis One of the easy ways of implementing multiplexing on an electrophoresis microchip is to facilitate channel parallelization, namely “microchip array electrophoresis” chips. Single-channel electrophoresis microchips have already shown clear advantages over the conventional slab gel or capillary electrophoresis with respect to analysis time, separation efficiency, and sample consumption, etc. However, microchips’ superiority will not be fully revealed without high sample throughput being realized on them. Fortunately, one of the most attractive features of the microfabricated devices is that intricate, ultra-dense channel arrays can be easily fabricated onto a single chip without significant increase in time and cost as compared to the single channel microchip fabrication. In 1997, Woolley et al. first demonstrated the feasibility of genotyping 12 samples simultaneously on a glass microchip with a rectilinear network of 12 channels [144]. Simpson et al. then expanded it to 48 channels with each channel capable of analyzing two samples sequentially. The system allowed an analysis of 96 samples in 8 minutes [145]. To further increase the density of the channel and address the limitation of a roster optical scanner, Shi et al. developed a rotary confocal fluorescence scanner together with a microchip that consists of 96 channels radially distributed on a 100mm-diameter wafer [146], which have been used to demonstrate a series of clinical genetic testing. The radial design was further advanced with the presentation of a 384-channel array on a 200mm-diameter wafer [147]. With 80 mm effective separation length, this microchip demonstrated genotyping of 384 individuals in only 325 s for the common hemochromatosis-linked H63D mutation in the human HFE gene, corresponding to a throughput of more than 1 sample per second. By multiplexing also in the detection aspect, for example spectral multiplexing, this format was expected to yield a throughput of 147,000 samples/day, a 20-fold increase over the commercial 96-capillary systems.
138 Zhu and Wang
While microchip array electrophoresis gained momentous successes in genotyping applications, several research teams, including Mathies’s group at UC Berkeley and Whitehead BioMEMS laboratory at MIT led by Ehrlich, were engaged in developing microchip based high throughput sequencing instruments in response to the call for “ultra-low-cost sequencing techniques” or “personal genome project”. The first DNA sequencing in a microchip array was demonstrated by Liu et al. in 2000 [148]. The 16 simple-cross channels were fanned out on a 100mm-diameter wafer. The system routinely yielded more than 450 bases in 15 min in each of the 16 channels. At almost the same time, Simpson et al. and Backhouse et al. both fabricated 48 straight channels on glass substrates [149, 150] for DNA sequencing. These micro channel array designs are a complete miniaturized version of a conventional capillary array system, on which electrokinetic sample injection along the separation channel rather than the cross-T injector was used. The typical advantages associated with the microchip, such as the shorter separation length and the faster analysis, were not fully realized in these works. Based on the theoretical prediction and experimental investigation on the turn geometries for minimizing band broadening in microchip electrophoresis [54, 151, 152], Paegel et al. introduced a “pinched turn” to extend the separation length from 55 mm to 159 mm on a 150mm-diameter wafer, which is necessary in DNA sequencing for single base resolution. An average read length of 430 bases at 99% base-call accuracy was obtained in 24 min, equivalent to a sequencing rate of 30 bases per second [153]. By replacing the LPA sieving matrix with Nanogel, a read length of about 500 bp at 98% calling accuracy was obtained in 25 minutes [92]. Researchers at Whitehead BioMEMS lab did a systematic study on the important parameters that affect the DNA sequencing performance in microchips and fabricated 32 identical 40-cm long microchannels on a 25 cm x 50 cm glass plate. This device achieved a read length of 800 bases, but at the cost of a longer separation time (80 min) [154-155]. While the race of maximizing the array density continues, the final trophy went to the prototype of BioMEMS-768 high throughput sequencing system [31, 156]. Two 384-channel plates were cycled alternatively between electrophoresis and regeneration. In each run, a total of greater than 172,000 bases, 99% calling accuracy was generated in ~70 minutes, corresponding to a total of a 4 megabase throughtput per day. For comparison, it takes about an hour for the latest ABI 3730xl 96-capillary DNA Sequencer to generate 67200 bases (700 bases each capillary) at an accuracy of 99% in a standard opera-
Genetic Analysis in Miniaturized Electrophoresis Systems 139
tion condition, corresponding to ~1.6 megabase per day. With the significant reduction in reagent consumption, microchip electrophoresis becomes one of the most promising technologies to deliver reasonably low cost whole genome sequencing (~$100,000 genome). A summary of multiplexed electrophoresis microchips is given in Table 4.1. Table 4.1 Summary of multichannel electrophoresis microchips for DNA sequencing Refer- # of Pattern ence Channels
Separa- Separa- Sample tion tion /Template Length Matrix (mm)
Accu- Read Time Length (min) racy (%) (bp)
[148]
16
Double T 70-76
LPA
Unknown
99
450
15
[150]
48
Straight
456
POP-6
BigDye Sample
98
640
150
[149]
48
Straight
100
LPA
M13mp18 97
400
[155]
32
Cross
400
LPA
M13mp18 98
800
78
[153]
96
Double T 159
LPA
M13mp18 99
430
24
[92]
96
Double T 159
NanoM13mp18 98.5 gel
500
24
[31]
768
Cross
M13mp18 98.5
580
~70
370450
LPA
4.4 Integration in Microchip Electrophoresis for Genetic Analysis 4.4.1 Sample preparation on microchip Sample preparation steps prior to electrophoresis, which include DNA extraction, amplification, and purification etc., are extremely critical for high
140 Zhu and Wang
quality genotyping or sequencing data. The conventional sample preparation procedures generally involve a series of benchtop instruments doing off-chip processing and the sample volumes involved are usually several orders of magnitude larger than what is needed for the application on a microchip. Those processing procedures are usually cumbersome and often limit the quoted advantages of microchip electrophoresis, such as low reagents consumption, low cost and high throughput. Novel methods for onchip sample preparation with the intention of integration with the downstream microchip application have been proposed and demonstrated. Sequence specific DNA fragment enrichment is generally required in genetic analysis. Since its inception, PCR (polymerase chain reaction) is the most powerful tool to amplify the target DNA fragment millions to billions of times. Implementing PCR on a microdevice has been of great interests to scientists and engineers. Microchip PCR devices can be considered having two categories: stationary-chamber PCR and continuous flow PCR. The stationary chamber PCR device is basically a miniaturized thermocycler. Nanoliter scale PCR reactions have been demonstrated in this format [157]. The continuous flow format, which limits the thermal mass to the sample solution instead of the whole devices, generally provides faster PCR [158, 159]. A few recent reviews have summarized the progresses of microchip PCR [160-163]. This chapter will only mention the PCR microdevices that are or were intended to be coupled to microchip electrophoresis. Mathies group developed a monolithic, nanoliter DNA amplification and electrophoresis system with a microfabricated thin film heater and resistive temperature detector (RTD) [164-166]. Landers group demonstrated a fast non-contact PCR reactor by employing an IR temperature control system [167, 168]. In DNA sequencing, amplification of various lengths of DNA fragments is typically carried out by a Sanger cycle sequencing reaction, a similar process as PCR with the difference being the random termination of the chain extension when ddNTP incorporated. Therefore, any microchip design that is suitable for PCR can be applied to Sanger sequencing reaction as well. Both stationary and continuous flow formats have been implemented for Sanger cycle sequencing reactions. The Mathies group used the abovementioned configuration to complete the Sanger extension on a microchip [60]. Soper et al. demonstrated Sanger sequencing with a 62 nL reactor coupled directly to a capillary gel column for DNA separation [169]. Compared to PCR, Sanger extension is much slower because the polymerase is altered to improve the incorporation efficiency of the phosphate analogs, e.g. ddNTPs or dye labeled ddNTPs. By improving the heat transfer with a
Genetic Analysis in Miniaturized Electrophoresis Systems 141
continuous flow microchip, Wang et al. pushed the Sanger reaction speed to the enzymatic limit [170]. A purification step is generally required before the PCR amplification of DNA from a crude biological sample or after the Sanger cycle sequencing reactions in a sequencing experiment. The purification before PCR amplification is to remove the inhibitors and/or pre-concentrate the DNA templates to ensure the efficiency of PCR. The purification of Sanger extension reaction products is routinely carried out before loading the sample onto a capillary or a microchip to remove the excess amount of DNA template, unincorporated nucleotides, and salts. Those impurities with high electrophoretic mobility could adversely affect the sequencing results including the read-length and the separation resolution. The conventional DNA purification methods are ethanol precipitation or solid phase extraction. Those are time-consuming and involve multiple steps that are not feasible to be incorporated on a microchip. Alternative approaches have been investigated for better purification quality and more importantly for an easy integration into a micro-total-analysis system. The Landers group explored solid-phase extraction (SPE), which relies on DNA adsorption to silica resin in the presence of a chaotropic agent, to extract, purify and pre-concentrate DNA from raw samples from human, bacteria, and virus etc. [171-173]. Xu et al. purified the Sanger extension fragments in a UV modified polycarbonate microchip with an adapted solid-phase reversible immobilization (SPRI) chemistry [174], which was first demonstrated with carboxylated magnetic microbeads by Hawkins et al [175, 176]. Paegel et al. demonstrated a low viscosity hydrogel matrix for selectively hybridizing and releasing the Sanger extension products to achieve the sample pre-concentration and purification [177]. 4.4.2 System integration Developing a fully integrated DNA analysis system on a single chip has been the ultimate goal of many researchers since the inception of the concept: lab-on-a-chip [178]. Typical sample transfer methods using conventional pipette tips or automatic sample dispensers at microliter scale obviously degrade most of the advantages associated with the microchip. The true power of the miniaturization will not be fully revealed without system integration. Although the microchip electrophoresis device alone has shown superior separation performance and the related sample preparation units have also shown great promise in processing nanoliter volumes of
142 Zhu and Wang
sample, integrating various functions together and streamlining the total operation on microchips still represents great challenges to researchers across all disciplines. Thus far, only a handful of integrated microchip genetic analyses have been demonstrated. The first integration, demonstrated by Burns et al. in 1998, reported an integrated nanoliter DNA analysis device that was capable of performing sample injection, amplification or digestion, and electrophoretic separation [179]. The liquid manipulation units, thermal control units, in-situ crosslinked polyacrylamide gel electrophoresis units, and photodiode detectors were all fabricated on a silicon substrate by conventional micromachining techniques (Fig. 4.5). One of the keys to the success of the proof-ofconcept device was the hydrophobic patch valves and capillary force pumps for nanoliter scale liquid manipulation. However, it is challenging for surface tension based pumps and valves to manipulate different fluids (e.g. real biological samples, different buffers) in more complex geometries.
Fig. 4.5 (Top) Schematic of integrated device with two liquid samples and electrophoresis gel present. (Bottom) Optical micrograph of the device from above. Wire bonds to the printed circuit board can be seen along the top edge of the device. The pressure manifold and buffer wells that fit over the entry holes at each end of the device are not shown. (From [179]. Reprinted with permission from AAAS.)
Genetic Analysis in Miniaturized Electrophoresis Systems 143
Following the first integrated “Lab-on-a-chip” device, there have been a few other attempts of integrating the sample preparation units with the electrophoresis units. The progress was partially hindered by the lack of interconnecting components, e.g. micropumps, micro-valves, and micromixers, for seamlessly interfacing various functional units and assembling them into a fully functionized and miniaturized system. By utilizing the pneumatically activated valves and pumps, two research protoytpes of miniaturized and fully integrated electrophoretic systems targeting real biological samples were finally revealed in 2006, nearly one decade after the first proof-of-concept lab-on-a-chip device. Mathies et al. incorporated all three Sanger sequencing steps, thermal cycling, purification, and electrophoretic separation, into a fully integrated and automated format (Fig. 4.6) [60]. With a 30-cm long separation channel filled by linear polyacrylamide gel, a read length of 556 bases with 99% accuracy from 1 fmole DNA template was successfully demonstrated. The reagent consumption and the cost of the analysis, was reduced by ~100 fold (from 20 µL to 250 nL), a dramatic improvement simply because of the removal of the off-chip sample transfer steps. The multi-layer glass-PDMS hybrid lab-on-a-chip assembly involved using several generic components that could perform the essential individual process, so it could be potentially used in a wide range of bioanalytical and biomedical applications. Compared to the pure DNA template used by Mathies et al., Landers et al. demonstrated a “sample-in-answer-out” device which was capable of accepting crude biological samples, such as blood and other bodily fluids [180]. Two sample preparation units, including a solid phase DNA extraction (SPE) unit to carry out on-chip DNA purification and a PCR amplification unit, were integrated to an electrophoretic analysis unit through a combination of several recent innovations in microfluidics, such as differential flow resistances, elastomeric valves, and laminar flows. The integrated microsystem could screen pathogen infection from a very small (sub-microliter) volume of crude samples in less than 30 minutes, as demonstrated by the identification of the presence of Bacillus anthracis (anthrax) in 0.75 µL of whole blood from infected but asymptomatic mice, and of Bordetella pertussis in 1 µL of nasal aspirate from a suspected human patient.
144 Zhu and Wang
Fig. 4.6 Integrated DNA sequencing bioprocessor. (Reprinted with permission from [60]. Copyright (2006) National Academy of Sciences, U.S.A.)
It has been a long way from the birth of the concept of “lab-on-a-chip” to the first published successful research prototypes. It may take another few years before the turnkey “sample-to-answer” micro-total analysis system eventually comes to the real world to be used for point-of-care clinical diagnostics or at the crime scene for forensic analysis. With the continuing growth of microfluidic technology and little expectation for such a technology to reach a plateau, scientists and engineers are taking on the daunting challenges and getting closer and closer to reaching their ultimate goal.
4.5 Commercial Microfluidic Instruments for Genetic Analyses An ultimate validation of a scientific idea may be its realization into a commercial product that subsequently benefits society. Since the conceptual birth of “lab-on-a-chip” almost two decades ago, there has been tre-
Genetic Analysis in Miniaturized Electrophoresis Systems 145
mendous interest and incredible technological advancement over the years, as evidenced by the enormous amount of publications partially summarized in the previous sections. Following its success at the research level and considering the great promise it holds for almost every life science and health care area, microfluidics technology is also expected to have widespread commercial use. The main drivers that promote microfluidics products into the commercial markets are the inherent merits of miniaturization mentioned previously: rapid analysis time, reduced sample volume and reagents cost, and great labor savings offered by system automation and integration. The reduction of assay time from hours to minutes or even seconds by using the microfluidics technology is the booster of the emerging microfluidics market. The micro, nano, or pico liter of volume needed for analysis and the same scale of waste produced on a microchip system is another market driving force. The cost per assay is dramatically reduced by ten or more fold due to the lower reagent consumption. In addition, its flexibility of interfacing with other methods and technologies for automation allows a total analysis to be carried out without user intervention, resulting in a great labor savings. The goal of this section is to summarize the major commercial microfluidic products on the market for genetic analysis and to give a general overview of the rising microfluidic markets. The authors have found that the microfluidics products evolve very rapidly with the changing market. Therefore, rather than concentrating on specific products, we will guide the readers through different companies by their unique technologies and then introduce their products based on such technologies. 4.5.1 Commercial microchip electrophoresis instruments for genetic analysis A summary of commercial microfluidic instruments that specialize in microchip electrophoresis will be first given in this section followed by a few other microfluidic products that carry out a broader scope of genetic analyses including sample preparation and micro-array applications. A few examples of companies who are making microchip electrophoresis products are: Agilent Technologies, Bio-Rad Laboratories, and Caliper Life Sciences. Their current products are 2100 Bioanalyzer, Experion Automated Electrophoresis System, and Labchip 90, respectively. All
146 Zhu and Wang
three products are using the same LapChip® technology developed by Caliper Life Sciences. The LapChip® technology was described by Bousse etc. [181]. Briefly, it automates steps involved in gel-based electrophoresis for separation of biomolecules including DNA, RNA, and protein. Electrophoretic separation, analyte staining and detection are integrated on a single chip and the process is automatically repeated for multiple-sample analysis. DNA sizing and quantification of ten samples can be completed in 30 min on a chip, a dramatic improvement over conventional capillary instruments. Agilent 2100 Bioanalyzer was the first (in 1999) commercialized microfluidic product based on Caliper’s LapChip® technology through collaboration between the two companies [182, 183]. Similarly, Bio-Rad launched the Experion Automated Electrophoresis System in 2004 using the same technology by collaborating with Caliper as well [183]. Bioanalyzer and Experion are considered as low-throughput miniaturized electrophoresis systems. The number of samples that can be analyzed depend on the number of wells on a chip and both systems currently analyze ten DNA samples per chip. Caliper’s own microchip electrophoresis product, LabChip90, is a high-throughput alternative to the Bioanalyzer and Experion systems. Different from the other two automated electrophoresis systems, LabChip90 can be coupled with 96 or 384 micro-titer plates, automatically draw samples from a well-plate sequentially using Caliper’s proprietary “sipper” sampling system, then analyze each sample under a typical gelelectrophoresis condition (www.caliperls.com). The system automates electrophoresis, detection, and data analysis for processing thousands of samples per day. A variety of chips and reagents designed for different applications are offered along with the instrument as consumables. The applications for DNA analysis include genotyping, micro-array production, and gene expression studies. The commonality of the above mentioned LapChip® technology-based microfluidic instruments are that they all integrate and automate certain functionality such as separation, staining, destaining, and detection in the workflow. However, the cumbersome sample preparation steps involved in a complete DNA analysis (e.g. from crude samples) are all done manually prior to sample loading onto the instruments. Those products are similar to the conventional bench top capillary or slab gel systems as to the application, but they provide faster analysis speed and higher throughput. These products have been well received by the customers and the success of these microfluidic products has helped the companies gaining big market shares in the microfluidic industry. The revolutionary separation speed
Genetic Analysis in Miniaturized Electrophoresis Systems 147
and data quality of the first microfluidic instruments on the market are evident enough for researchers to realize the power of the new generation of separation platform. 4.5.2 Integrated microfluidic instruments for genetic analyses The possibility of the many steps of conventional assays being streamlined into a single process and in the nanoliter scale is very attractive. An ideal lab-on-a-chip system should automate and integrate all steps involved in an analysis on one platform. The “sample-in” and “result-out” format of Lab-on-a-chip or Micro-total-analysis system reduces analysis time, risk of sample loss or contamination, and is a perfect replacement for bulky, expensive laboratory robots. In this section, we introduce a few companies who are developing integrated genetic analysis instruments using their proprietary microfluidics technologies. Cepheid is one of the pioneers in commercializing the integrated genetic analysis system (www.cepheid.com). By using its smart cycler technology (I-Core®) that integrates thermal cycling with micro-optics and electronics, Cepheid offered customers two versions of PCR-based genetic analysis systems: the SmartCycler® and the GeneXpert®. These products have been used in rapid genetic screening for infectious diseases, cancer, and bio-terrorism detection. Cepheid offers series of testing and reagent kits to go with the instruments to amplify and identify a specific DNA sequence associated with a disease or bio-threat agents. SmartCycler® is a real-time PCR system that can provide answers from a processed biological sample in 30-40 minutes. The GeneXpert® system integrates sample preparation with PCR amplification and real time detection for a fully automated genetic analysis. It purifies, concentrates, amplifies and detects targeted nucleic acid sequences in less than 30 minutes from a raw sample. It uses a disposable microfluidic cartridge together with a special designed PCR reaction tube. After raw specimens lysed by ultrasonic force, the released DNA and reaction mixtures are pumped through chambers inside the cartridge by a syringe drive and a rotary drive and finally reach the reaction tube for thermal cycling and detection. The throughput of the system can be scalable up to 16 modules at this moment with each module carrying out one test (www.cepheid.com). Microchip Biotechnologies (MBI) is another microfluidic company that is developing a “sample-to-answer” format of genetic analysis system (http://www.microchipbiotech.com). Specifically, they are developing an
148 Zhu and Wang
integrated microfluidic system that includes sample preparation and analysis for DNA sequencing. The two key technologies: Microbead Capture technology and Microscale on-chip valves (MOV™) technology focus on the interfacing between the macro- and micro- world. The Microbead Capture technology uses coated magnetic beads to capture and concentrate specific biological molecules from a large amount of complexed sample matrices into micro- or nanoliter size. It serves to bridge large volume sample with the downstream micro-volume analysis. MOV™ technology is one of the on-chip “valving and pumping” designs that enable the successful micro or nanofluidic mixing and fluid transferring within the intricate microfluidic network. The on-chip valves and pumps are the key components that allow system miniaturization, automation, and integration. Combining these two key technologies and together with a robotic workstation, MBI has developed an automated sequencing sample preparation instrument, Apollo 100. This instrument automatically sets up on-chip Sanger cycle sequencing reaction and then perform sample clean up for the downstream detection. During the automated process, robots move the DNA template and the reaction mix onto the microchip, MOV™ assist with pumping and mixing of samples for thermocycling and finally the magnetic beads come into place to cleanup the sequencing reaction products. Not connected with a DNA sequencer and also using pre-processed DNA template, the system is not fully automated in terms of sequencing from a raw sample. However, the Apollo 100 will be beneficial to those who already own capillary DNA sequencers but perform cumbersome sample preparation steps manually (http://www.microchipbiotech.com).. Fluidigm introduced the technology of “integrated fluidic circuits (IFCs)” to the microfluidic industry and launched the BioMark™ product line based on such a technology for variety of genetic analysis (www.fluidigm.com). IFCs are microdevices fabricated with networks of fluid-control valves (NanoFlex™ valve) and interconnected channels, in which the movement of the biological sample and reagents can be regulated. The core of this technology is the NanoFlex™ valve which consists of layers of elastomers that deflect under pressure to form a seal. Using NanoFlex™ valves’ compact size and precise fluidic control, Fluidigm established high throughput genetic assay platforms, the Biomark™ 48.48 Dynamic Array and Biomark™ digital arrays. The Biomark™ products offer real-time PCR and genotyping at a high throughput together with accurate DNA quantification (www.fluidigm.com). If microscale liquid handling and micro-array printing are also considered within the realm of the microfluidics, Akonni Biosysem may need to be
Genetic Analysis in Miniaturized Electrophoresis Systems 149
mentioned for its TruArray® technology (www.akonni.com). It is an integrated microarray technology that screens a biological sample against hundreds of diseases markers simultaneously and gives an answer in less than 30 min. An array of the micro-gel-drops is precisely loaded onto a plastic microfluidic slide, size of a credit card. Each micro-gel-drop carries out an individual PCR reaction, tailored for a specific diagnostic test. Coupling this on-chip PCR with precise microfluidic control, the portable TruArray® system enables hundreds to thousands of diagnostic tests performed on a clinical sample in a short period of time (www.akonni.com). There are also dozens of other companies that hold patented or licensed microfluidics technologies with their products spanning over a broad spectrum of applications. A table (Table 4.2) that summarizes the current active microfluidic companies and their major technologies for genetic analysis is given below including some examples that have not been discussed in the text. The authors do not intended to provide a complete list of all the microfluidics companies and all their microfluidic products. Even at the time of this book being published, dozens of new companies are getting started and new product ideas being introduced. Table 4.2 Some microfluidic companies and their commercialized microfluidic instruments for genetic analysis Company
Akonni Biosystems www.akonni.com Agilent Technologies www.agilent.com Bio-Rad Laboratories www.bio-rad.com
Technology TruArray™ with microfluidic liquid handling) LabChip® for gel electrophoresis LabChip® for gel electrophoresis
Genetic tion
Applica-
Clinical molecular diagnosis Electrophoresis and quantification Electrophoresis and quantification n
Bio-Trove www.biotrove.com
OpenArray™
Genotyping, time PCR
Caliper Life Sciences www.caliperls.com
LabChip® for gel electrophoresis
Cepheid www.cepheid.com
Smart Cycler technology (I-Core®)
Electrophoresis and quantification, Sequencing Real-time PCR, Clinical diagnosis
real-
Product for genetic analysis TruArray™ 2100 Bioanalyzer Experion Automated Electrophoresis System OpenAray™ SNP Genotyping System LabChip® 90
GeneXpert SmartCycler
150 Zhu and Wang Table 4.2 cont. Fluidigm www.fluidigm.com
Integrated Fluidic circuits (IFCs) NanoFlex™ valve\
Handylab www.handylab.com
Microthermal Circuit
Microchip Biotechnologies www.microchipbiotec h.com Nanogen www.nanogen.com
Microbead Capture Technology, Microscale on-chip valves (MOV™) Nanochip® Electronic Microarray
Genotyping, gene expression profiling, quantification, diagnostic Real-time PCR, Clinical diagnosis DNA Sequencing
SNP analysis, STR analysis, mutation analysis
BioMark™ Dynamic array BioMark™ Digital array Jajuar®, Lynx® (under development) Apollo 100
NanoChip® 400 system (product closed in 2007)
4.6 Microfluidic Markets and Future Perspectives The market for microfluidics/lab-on-a-chip products is growing rapidly with the technologies developed in the field. With the significant progress and the high expectation in both the technology and the market, customers have begun to accept the microfluidic products . One of the hurdles the customer has to overcome before the microfluidic products can stand firmly in the market is the replacement of existing well-functioning and highly invested large-scale robotic systems with the maybe even more expensive microfluidic products. In addition, products that come out of new technologies often have risks associated with them, even though many considerations are put into improving the products’ reliability and cost. However, with a considerable amount of investment from both government and private sectors continuing to go into the R&D laboratories of numerous start-up and well-established companies to develop microfluidic instruments, the market potential for microfluidic industry seems to be rising. More and more customers may gradually realize the long-term savings effect brought by the microfluidic technology and decide to invest in it, especially in the cases where reduction of cost and limited amount of sample are prominent. In the meantime, many microfluidics manufacturers are striving for greater accuracy, improved throughput, increased automation, and better robustness. There seems to be quite potential market spaces for this continuing maturing technology. Many are expecting the microfluidic
Genetic Analysis in Miniaturized Electrophoresis Systems 151
market to flourish in the next a few years as the novel technology gains more acceptances among the research labs around the world. The continued development of microfluidic industry, on the other hand, will depend upon the successful realization of miniaturization and integration. Performing all steps of a biological assay on a single microchip is very attractive for its obvious advantages in terms of speed, cost, and automation. Currently, a lot of robotic workstations are being used as enabling tools at the interface of microchips and sample processing steps. Although they have provided great productivity and accuracy, such complicated and expensive equipment may not hold up long under the increasing demands for portable and fieldable integrated total DNA analysis systems. For that reason, the microfluidics industry may aim at creating new genetic analysis platforms that incorporate greater functionalities but still with a small footprint. As indicated by the latest research publications, as well as a few products listed above, one of the obvious research trends is to produce a “sample-in, answer out” high-capacity microfluidic product. Microfluidics manufactures are gaining momentum through continuous technological breakthroughs that are able to address the unmet market needs. The gap between where the technology is and where the market needs it to be continues to get smaller and smaller.
References 1. Lander ES, et al. (2001) Initial sequencing and analysis of the human genome. Nature 409:860-921. 2. Venter JC, et al. (2001) The sequence of the human genome. Science 291: 1304-1351. 3. Shendure J, Mitra RD, Varma C, Church GM (2004) Advanced sequencing technologies: Methods and goals. Nature Reviews Genetics 5:335-344. 4. Chan EY (2005) Advances in sequencing technology. Mutation ResearchFundamental and Molecular Mechanisms of Mutagenesis 573:13-40. 5. Metzker ML (2005) Emerging technologies in DNA sequencing. Genome Research 15:1767-1776. 6. Fredlake CP, Hert DG, Mardis ER, Barron AE (2006) What is the future of electrophoresis in large-scale genomic sequencing? Electrophoresis 27:36893702. 7. DeRisi J, Penland L, Brown PO, Bittner ML, Meltzer PS, Ray M, Chen YD, Su YA, Trent JM (1996) Use of a cDNA microarray to analyse gene expression patterns in human cancer. Nature Genetics 14:457-460. 8. Fodor SPA, Rava RP, Huang XHC, Pease AC, Holmes CP, Adams CL (1993) Multiplexed biochemical assays with biological chips. Nature 364:555-556.
152 Zhu and Wang 9. Fodor SPA, Read JL, Pirrung MC, Stryer L, Lu AT, Solas D (1991) Lightdirected, spatially addressable parallel chemical synthesis. Science 251:767773. 10. Pease AC, Solas D, Sullivan EJ, Cronin MT, Holmes CP, Fodor SPA (1994) Light-generated oligonucleotide arrays for rapid dna-sequence analysis. Proceedings of the National Academy of Sciences of the United States of America 91:5022-5026. 11. Schena M, Shalon D, Davis RW, Brown PO (1995) Quantitative monitoring of gene-expression patterns with a complementary-dna microarray. Science 270:467-470. 12. Shalon D, Smith SJ, Brown PO (1996) A DNA microarray system for analyzing complex DNA samples using two-color fluorescent probe hybridization. Genome Research 6:639-645. 13. Churchill GA (2002) Fundamentals of experimental design for cDNA microarrays. Nature Genetics 32:490-495. 14. Gunderson KL, Steemers FJ, Lee G, Mendoza LG, Chee MS (2005) A genome-wide scalable SNP genotyping assay using microarray technology. Nature Genetics 37:549-554. 15. Heller MJ (2002) DNA microarray technology: Devices, systems, and applications. Annual Review of Biomedical Engineering 4:129-153. 16. Hoheisel JD (2006) Microarray technology: beyond transcript profiling and genotype analysis. Nature Reviews Genetics 7:200-210. 17. Holloway AJ, van Laar RK, Tothill RW, Bowtell DDL (2002) Options available - from start to finish - for obtaining data from DNA microarrays. Nature Genetics 32:481-489. 18. Pinkel D, Albertson DG (2005) Array comparative genomic hybridization and its applications in cancer. Nature Genetics 37:S11-S17. 19. Slonim DK (2002) From patterns to pathways: gene expression data analysis comes of age. Nature Genetics 32:502-508. 20. Syvanen AC (2005) Toward genome-wide SNP genotyping. Nature Genetics 37:S5-S10. 21. Venkatasubbarao S (2004) Microarrays - status and prospects. Trends in Biotechnology 22:630-637. 22. Dolnik V, Liu SR (2005) Applications of capillary electrophoresis on microchip. Journal of Separation Science 28:1994-2009. 23. Horsman KM, Bienvenue JM, Blasier KR, Landers JP (2007) Forensic DNA analysis on microfluidic devices: A review. Journal of Forensic Sciences 52:784-799. 24. Jin LJ, Ferrance J, Landers JP (2001) Miniaturized electrophoresis: An evolving role in laboratory medicine. Biotechniques 31:1332-1335. 25. Kan CW, Fredlake CP, Doherty EAS, Barron AE (2004) DNA sequencing and genotyping in miniaturized electrophoresis systems. Electrophores3564-3588. 26. Oin JH, Fung YS, Lin BC (2003) DNA diagnosis by capillary electrophoresis and microfabricated electrophoretic devices. Expert Review of Molecular Diagnostics 3:387-394.
Genetic Analysis in Miniaturized Electrophoresis Systems 153 27. Sinville R, Soper SA (2007) High resolution DNA separations using microchip electrophoresis. Journal of Separation Science 30:1714-1728. 28. Szantai E, Guttman A (2006) Genotyping with microfluidic devices. Electrophoresis 27:4896-4903. 29. Verpoorte E (2002) Microfluidic chips for clinical and forensic analysis. Electrophoresis 23:677-712. 30. Sanger F, Nicklen S, Coulson AR (1977) DNA sequencing with chainterminating inhibitors. Proceedings of the National Academy of Sciences of the United States of America 74:5463-5467. 31. Aborn JH, El-Difrawy SA, Novotny M, Gismondi EA, Lam R, Matsudaira P, McKenna BK, O'Neil T, Streechon P, Ehrlich DJ (2005) A 768-lane microfabricated system for high-throughput DNA sequencing. Lab on A Chip 5:669-674. 32. Llopis SL, Osiri J, Soper SA (2007) Surface modification of poly(methyl methacrylate) microfluidic devices for high-resolution separations of singlestranded DNA. Electrophoresis 28:984-993. 33. Shi YN (2006) DNA sequencing and multiplex STR analysis on plastic microfluidic devices. Electrophoresis 27:3703-3711. 34. Altshuler D, Brooks LD, Chakravarti A, Collins FS, Daly MJ, Donnelly P, Int HapMap, C (2005) A haplotype map of the human genome. Nature 437:12991320. 35. Feuk L, Carson AR, Scherer SW (2006) Structural variation in the human genome. Nature Reviews Genetics 7:85-97. 36. Syvanen AC (2001) Accessing genetic variation: Genotyping single nucleotide polymorphisms. Nature Reviews Genetics 2:930-942. 37. Hestekin CN, Barron AE (2006) The potential of electrophoretic mobility shift assays for clinical mutation detection. Electrophoresis 27:3805-3815. 38. Guttman A, Gao HG, Haas R (2001) Rapid analysis of mitochondrial DNA heteroplasmy in diabetes by gel-microchip electrophoresis. Clinical Chemistry 47:1469-1472. 39. Qin JH, Liu ZY, Wu DP, Zhu N, Zhou XM, Fung YS, Lin BC (2005) Genotyping the-6A/G functional polymorphism in the core promoter region of angiotensinogen gene by microchip electrophoresis. Electrophoresis 26:219-224. 40. Tian HJ, Jaquins-Gerstl A, Munro N, Trucco M, Brody LC, Landers JP (2000) Single-strand conformation polymorphism analysis by capillary and microchip electrophoresis: A fast, simple method for detection of common mutations in BRCA1 and BRCA2. Genomics 63:25-34. 41. Tian HJ, Emrich CA, Scherer JR, Mathies RA, Andersen PS, Larsen LA, Christiansen M (2005) High-throughput single-strand conformation polymorphism analysis on a microfabricated capillary array electrophoresis device. Electrophoresis 26:1834-1842. 42. Tian HJ, Brody LC, Landers JP (2000) Rapid detection of deletion, insertion, and substitution mutations via heteroduplex analysis using capillary- and microchip-based electrophoresis. Genome Research 10:1403-1413. 43. Kourkine IV, Hestekin CN, Buchholz BA, Barron AE (2002) Highthroughput, high-sensitivity genetic mutation detection by tandem single-
154 Zhu and Wang strand conformation polymorphism/heteroduplex analysis capillary array electrophoresis. Anal Chem 74:2565-2572. 44. Manage DP, Zheng Y, Somerville MJ, Backhouse CJ (2005) On-chip HA/SSCP for the detection of hereditary haemochromatosis. Microfluidics and Nanofluidics 1:364-372. 45. Tian HJ, Brody LC, Fan SJ, Huang ZL, Landers JP (2001) Capillary and microchip electrophoresis for rapid detection of known mutations by combining allele-specific DNA amplification with heteroduplex analysis. Clinical Chemistry 47: 73-185. 46. Vahedi G, Kaler C, Backhouse CJ (2004) An integrated method for mutation detection using on-chip sample preparation, single-stranded conformation polymorphism, and heterduplex analysis. Electrophoresis 25:2346-2356. 47. Butler JM (2006) Genetics and genomics of core short tandem repeat loci used in human identity testing. Journal of Forensic Sciences 51:253-265. 48. Jobling MA, Gill P (2004) Encoded evidence: DNA in forensic analysis. Nature Reviews Genetics 5:739-751. 49. Yeung SHI, Greenspoon SA, McGuckian A, Crouse CA, Emrich CA, Ban J, Mathies RA (2006) Rapid and high-throughput forensic short tandem repeat typing using a 96-lane microfabricated capillary array electrophoresis microdevice. Journal of Forensic Sciences 51:740-747. 50. Shi YN, Anderson RC (2003) High-resolution single-stranded DNA analysis on 4.5 cm plastic electrophoretic microchannels. Electrophoresis 24:33713377. 51. Harrison DJ, Fluri K, Seiler K, Fan ZH, Effenhauser CS, Manz A (1993) MICROMACHINING A miniaturized capillary electrophoresis-based chemical-analysis system on a chip. Science 261:895-897. 52. Harrison DJ, Manz A, Fan ZH, Ludi H, Widmer HM (1992) Capillary electrophoresis and sample injection systems integrated on a planar glass chip. Anal Chem 64:1926-1932. 53. Jacobson SC, Hergenroder R, Koutny LB, Ramsey JM (1994) High-speed separations on a microchip. Anal Chem 66:1114-1118. 54. Jacobson SC, Hergenroder R, Koutny LB, Warmack RJ, Ramsey JM (1994) Effects of injection schemes and column geometry on the performance of microchip electrophoresis devices. Anal Chem 66: 107-1113. 55. Jacobson SC, Moore AW, Ramsey JM (1995) FUSED QUARTZ SUBSTRATES FOR MICROCHIP ELECTROPHORESIS. Anal Chem 67:2059-2063. 56. Manz A, Fettinger JC, Verpoorte E, Ludi H, Widmer HM, Harrison DJ (1991) Micromachining of monocrystalline silicon and glass for chemical-analysis systems - a look into next century technology or just a fashionable craze. Trac-Trends in Analytical Chemistry 10:144-149. 57. Seiler K, Harrison DJ, Manz A (1993) Planar glass chips for capillary electrophoresis - repetitive sample injection, quantitation, and separation efficiency. Anal Chem 65:1481-1488.
Genetic Analysis in Miniaturized Electrophoresis Systems 155 58. McDonald JC, Duffy DC, Anderson JR, Chiu DT, Wu HK, Schueller OJA, Whitesides GM (2000) Fabrication of microfluidic systems in poly(dimethylsiloxane). Electrophoresis 21:27-40. 59. Sia SK, Whitesides GM (2003) Microfluidic devices fabricated in poly(dimethylsiloxane) for biological studies. Electrophoresis 24:3563-3576. 60. Blazej RG, Kumaresan P, Mathies RA (2006) Microfabricated bioprocessor for integrated nanoliter-scale Sanger DNA sequencing. Proceedings of the National Academy of Sciences of the United States of America 103:7240-7245. 61. Hong JW, Fujii T, Seki M, Yamamoto T, Endo I (2001) Integration of gene amplification and capillary gel electrophoresis on a polydimethylsiloxaneglass hybrid microchip. Electrophoresis 22:328-333. 62. Liu CC, Cui DF, Cai HY, Chen X, Geng ZX (2006) A rigid poly(dimethylsiloxane) sandwich electrophoresis microchip based on thincasting method. Electrophoresis 27: 2917-2923. 63. Becker H, Gartner C (2000) Polymer microfabrication methods for microfluidic analytical applications. Electrophoresis 21:12-26. 64. Becker H, Locascio LE (2002) Polymer microfluidic devices. Talanta 56:267287. 65. Fiorini GS, Chiu DT (2005) Disposable microfluidic devices: fabrication, function, and application. Biotechniques 38:429-446. 66. Rossier J, Reymond F, Michel PE (2002) Polymer microfluidic chips for electrochemical and biochemical analyses. Electrophoresis 23:858-867. 67. Szekely L, Guttman A (2006) Comparison of various channel fabrication methods for microchip electrophoresis. Current Analytical Chemistry 2:195201. 68. Pugmire DL, Waddell EA, Haasch R, Tarlov MJ, Locascio E (2002) Surface characterization of laser-ablated polymers used for microfluidics. Anal Chem 74:871-878. 69. McCormick RM, Nelson RJ, AlonsoAmigo MG, Benvegnu J, Hooper HH (1997) Microchannel electrophoretic separations of DNA in injection-molded plastic substrates. Anal Chem 69:2626-2630. 70. Lee GB, Chen SH, Huang GR, Sung WC, Lin YH (2001) Microfabricated plastic chips by hot embossing methods and their applications for DNA separation and detection. Sensors and Actuators B-Chemical 75:142-148. 71. Martynova L, Locascio LE, Gaitan M, Kramer GW, Christensen RG, MacCrehan WA (1997) Fabrication of plastic microfluid channels by imprinting methods. Anal Chem 69:4783-4789. 72. Heller C (2001) Principles of DNA separation with capillary electrophoresis. Electrophoresis 22:629-643. 73. Sartori A, Barbier V, Viovy JL (2003) Sieving mechanisms in polymeric matrices. Electrophoresis 24:421-440. 74. Slater GW, Guillouzic S, Gauthier MG, Mercier JF, Kenward M, McCormick LC, Tessier F (2002) Theory of DNA electrophoresis (similar to 19992002(1)/(2)). Electrophoresis 23:3791-3816.
156 Zhu and Wang 75. Slater GW, Kenward M, McCormick LC, Gauthier MG (2003) The theory of DNA separation by capillary electrophoresis. Current Opinion in Biotechnology 14:58-64. 76. Ogston AG (1958) The spaces in a uniform random suspension of fibres. Transactions of the Faraday Society 54:1754-1757. 77. Lerman LS, Frisch HL (1982) Why does the electrophoretic mobility of dna in gels vary with the length of the molecule. Biopolymers 21:995-997. 78. Lumpkin OJ, Dejardin P, Zimm BH (1985) Theory of gel-electrophoresis of dna. Biopolymers 24:1573-1593. 79. Barron AE, Sunada WM, Blanch HW (1995) The use of coated and uncoated capillaries for the electrophoretic separation of dna in dilute polymersolutions. Electrophoresis 16:64-74. 80. Barbier V, Viovy JL (2003) Advanced polymers for DNA separation. Current Opinion in Biotechnology 14:51-57. 81. Buchholz BA, Doherty EAS, Albarghouthi MN, Bogdan FM, Zahn JM, Barron AE (2001) Microchannel DNA sequencing matrices with a thermally controlled "viscosity switch". Anal Chem 73:157-164. 82. Buchholz BA, Shi W, Barron AE (2002) Microchannel DNA sequencing matrices with switchable viscosities. Electrophoresis 23:1398-1409. 83. Xu F, Baba Y (2004) Polymer solutions and entropic-based systems for double-stranded DNA capillary electrophoresis and microchip electrophoresis. Electrophoresis 25:2332-2345. 84. Hong JW, Hosokawa K, Fujii T, Seki M, Endo I (2001) Microfabricated polymer chip for capillary gel electrophoresis. Biotechnology Progress 17:958-962. 85. Ugaz VM, Lin RS, Srivastava N, Burke DT, Burns MA (2003) A versatile microfabricated platform for electrophoresis of double- and single-stranded DNA. Electrophoresis 24:151-157. 86. Zhao DS, Roy B, McCormick MT, Kuhr WG, Brazill SA (2003) Rapid fabrication of a poly(dimethylsiloxane) microfluidic capillary gel electrophoresis system utilizing high precision machining. Lab on A Chip 3:93-99. 87. Carrilho E, RuizMartinez MC, Berka J, Smirnov I, Goetzinger W, Miller AW, Brady D, Karger BL (1996) Rapid DNA sequencing of more than 1000 bases per run by capillary electrophoresis using replaceable linear polyacrylamide solutions. Anal. Chem. 68:3305-3313. 88. Pariat YF, Berka J, Heiger DN, Schmitt T, Vilenchik M, Cohen AS, Foret F, Karger BL (1993) Separation of dna fragments by capillary electrophoresis using replaceable linear polyacrylamide matrices. Journal of Chromatography A 652:57-66. 89. Salas-Solano O, Carrilho E, Kotler L, Miller AW, Goetzinger W, Sosic Z, Karger BL (1998) Routine DNA sequencing of 1000 bases in less than one hour by capillary electrophoresis with replaceable linear polyacrylamide solutions. Anal. Chem. 70:3996-4003. 90. Zhou HH, Miller AW, Sosic Z, Buchholz B, Barron AE, Kotler L, Karger BL (2000) DNA sequencing up to 1300 bases in two hours by capillary electro-
Genetic Analysis in Miniaturized Electrophoresis Systems 157 phoresis with mixed replaceable linear polyacrylamide solutions. Anal Chem 72:1045-1052. 91. Doherty EAS, Kan CW, Barron AE (2003) Sparsely cross-linked "nanogels" for microchannel DNA sequencing. Electrophoresis 24:4170-4180. 92. Doherty EAS, Kan CW, Paegel BM, Yeung SHI, Cao ST, Mathies RA, Barron AE (2004) Sparsely cross-linked "nanogel" matrixes as fluid, mechanically stabilized polymer networks for high-throughput microchannel DNA sequencing. Anal Chem 76:5249-5256. 93. Song LG, Liang DH, Fang DF, Chu B (2001) Fast DNA sequencing up to 1000 bases by capillary electrophoresis using poly(N,N-dimethylacrylamide) as a separation medium. Electrophoresis 22:1987-1996. 94. Song MJ, Lee DS, Ahn JH, Kim DJ, Kim SC (2004) Thermosensitive sol-gel transition behaviors of poly(ethylene oxide)/aliphatic polyester/poly (ethylene oxide) aqueous solutions. Journal of Polymer Science Part a-Polymer Chemistry 42:772-784. 95. Sudor J, Barbier V, Thirot S, Godfrin D, Hourdet D, Millequant R, Blanchard J, Viovy JL (2001) New block-copolymer thermoassociating matrices for DNA sequencing: Effect of molecular structure on rheology and resolution. Electrophoresis 22:720-728. 96. Kim Y, Yeung ES (1997) Separation of DNA sequencing fragments up to 1000 bases by using poly(ethylene oxide)-filled capillary electrophoresis. Journal of Chromatography A 781:315-325. 97. Wei W, Yeung ES (2000) Improvements in DNA sequencing by capillary electrophoresis at elevated temperature using poly(ethylene oxide) as a sieving matrix. Journal of Chromatography B 745:221-230. 98. Xu F, Jabasini M, Baba Y (2005) Screening of mixed poly(ethylene oxide) solutions for microchip separation of double-stranded DNA using an orthogonal design approach. Electrophoresis 26:3013-3020. 99. Gao QF, Pang HM, Yeung ES (1999) Simultaneous genetic typing from multiple short tandem repeat loci using a 96-capillary array electrophoresis system. Electrophoresis 20:1518-1526. 100. Gao QF, Yeung ES (1998) A matrix for DNA separation: Genotyping and sequencing using poly(vinylpyrrolidone) solution in uncoated capillaries. Anal Chem 70:1382-1388. 101. Kim DK, Kang SH (2005) On-channel base stacking in microchip capillary gel electrophoresis for high-sensitivity DNA fragment analysis. Journal of Chromatography A 1064:121-127. 102. Song JM. Yeung ES (2001) Optimization of DNA electrophoretic behavior in poly(vinyl pyrrolidone) sieving matrix for DNA sequencing. Electrophoresis 22:748-754. 103. Moritani T, Yoon K, Rafailovich M, Chu B (2003) DNA capillary electrophoresis using poly(vinyl alcohol). I. Inner capillary coating. Electrophoresis 24:2764-2771. 104. Haynes RD, Meagher RJ, Won JI, Bogdan FM, Barron AE (2005) Comblike, monodisperse polypeptoid drag-tags for DNA separations by end-labeled freesolution electrophoresis (ELFSE). Bioconjugate Chemistry 16:929-938.
158 Zhu and Wang 105. Won JI (2006) Recent advances in DNA sequencing by End-Labeled FreeSolution Electrophoresis (ELFSE). Biotechnology and Bioprocess Engineering 11:179-186. 106. Volkmuth WD, Austin RH (1992) Dna Electrophoresis in Microlithographic Arrays. Nature 358:600-602. 107. Volkmuth WD, Duke T, Wu MC, Austin RH, Szabo A (1994) Dna Electrodiffusion in A 2D Array of Posts. Physical Review Letters 72:2117-2120. 108. Duke TAJ, Austin RH, Cox EC, Chan SS (1996) Pulsed-field electrophoresis in microlithographic arrays. Electrophoresis 17:1075-1079. 109. Duke TAJ, Austin RH (1998) Microfabricated sieve for the continuous sorting of macromolecules. Physical Review Letters 80:1552-1555. 110. Huang LR, Tegenfeldt JO, Kraeft JJ, Sturm JC, Austin RH, Cox EC (2002) A DNA prism for high-speed continuous fractionation of large DNA molecules. Nature Biotechnology 20:1048-1051. 111. Kaji N, Tezuka Y, Takamura Y, Ueda M, Nishimoto T, Nakanishi H, Horiike Y, Baba Y (2004) Separation of long DNA molecules by quartz nanopillar chips under a direct current electric field. Anal Chem 76:15-22. 112. Doyle PS, Bibette J, Bancaud A, Viovy JL (2002) Self-assembled magnetic matrices for DNA separation chips. Science 295:2237-2237. 113. Minc N, Futterer C, Dorfman K, Bancaud A, Gosse C, Goubault C, Viovy JL (2004) Quantitative microfluidic separation of DNA in self-assembled magnetic matrixes. Anal. Chem. 76:3770-3776. 114. Minc N, Bokov P, Zeldovich KB, Futterer C, Viovy JL, Dorfman KD (2005) Motion of single long DNA molecules through arrays of magnetic columns. Electrophoresis 26:362-375. 115. Chou CF, Bakajin O, Turner SWP, Duke TAJ, Chan SS, Cox EC, Craighead HG, Austin RH (1999) Sorting by diffusion: An asymmetric obstacle course for continuous molecular separation. Proceedings of the National Academy of Sciences of the United States of America 96:13762-13765. 116. Cabodi M, Chen YF, Turner SWP, Craighead HG, Austin RH (2002) Continuous separation of biomolecules by the laterally asymmetric diffusion array with out-of-plane sample injection. Electrophoresis 23:3496-3503. 117. Huang LR, Cox EC, Austin RH, Sturm JC (2003) Tilted Brownian ratchet for DNA analysis. Anal Chem 75:6963-6967. 118. Hammond RW, Bader JS, Henck SA, Deem MW, McDermott GA, Bustillo JM, Rothberg JM (2000) Differential transport of DNA by a rectified Brownian motion device. Electrophoresis 21:74-80. 119. Bader JS, Hammond RW, Henck SA, Deem MW, McDermott GA, Bustillo JM, Simpson JW, Mulhern GT, Rothberg JM (1999) DNA transport by a micromachined Brownian ratchet device. Proceedings of the National Academy of Sciences of the United States of America 96:13165-13169. 120. Ros A, Eichhorn R, Regtmeier J, Duong TT, Reimann P, Anselmetti D (2005) Brownian motion - Absolute negative particle mobility. Nature 436:928-928. 121. Ros A, Hellmich W, Regtmeier J, Duong TT, Anselmetti D (2006) Bioanalysis in structured microfluidic systems. Electrophoresis 27:2651-2658.
Genetic Analysis in Miniaturized Electrophoresis Systems 159 122. Han J, Craighead HG (2000) Separation of long DNA molecules in a microfabricated entropic trap array. Science 288:1026-1029. 123. Cabodi M, Turner SWP, Craighead HG (2002) Entropic recoil separation of long DNA molecules. Anal Chem 74:5169-5174. 124. Turner SWP, Cabodi M, Craighead HG (2002) Confinement-induced entropic recoil of single DNA molecules in a nanofluidic structure. Physical Review Letters 88:128103. 125. Belder D, Ludwig M (2003) Surface modification in microchip electrophoresis. Electrophoresis 24: 3595-3606. 126. Dolnik V (2004) Wall coating for capillary electrophoresis on microchips. Electrophoresis 25:3589-3601. 127. Makamba H, Kim JH, Lim K, Park N, Hahn JH (2003) Surface modification of poly(dimethylsiloxane) microchannels. Electrophoresis 24:3607-3619. 128. Pallandre A, de Lambert B, Attia R, Jonas AM, Viovy JL (2006) Surface treatment and characterization: Perspectives to electrophoresis and lab-onchips. Electrophoresis 27:584-610. 129. Steiner F, Hassel M (2003) Control of electroosmotic flow in nonaqueous capillary electrophoresis by polymer capillary coatings. Electrophoresis 24:399-407. 130. Hjerten S (1985) High-performance electrophoresis - elimination of electroendosmosis and solute adsorption. Journal of Chromatography 347:191-198. 131. Schmalzing D, Adourian A, Koutny L, Ziaugra L, Matsudaira P, Ehrlich D (1998) DNA sequencing on microfabricated electrophoretic devices. Anal Chem 70:2303-2310. 132. Schmalzing D, Koutny L, Adourian A, Belgrader P, Matsudaira P, Ehrlich D (1997) DNA typing in thirty seconds with a microfabricated device. Proceedings of the National Academy of Sciences of the United States of America 94:10273-10278. 133. Tian HJ, Brody LC, Mao D, Landers JP (2000) Effective capillary electrophoresis-based heteroduplex analysis through optimization of surface coating and polymer networks. Anal Chem 72:5483-5492. 134. Effenhauser CS, Bruin GJM, Paulus A, Ehrat M (1997) Integrated capillary electrophoresis on flexible silicone microdevices: Analysis of DNA restriction fragments and detection of single DNA molecules on microchips. Anal Chem 69:3451-3457. 135. Han FT, Huynh BH, Ma YF, Lin BC (1999) High efficiency DNA separation by capillary electrophoresis in a polymer solution with ultralow viscosity. Anal Chem 71:2385-2389. 136. Lee GB, Lin CH, Lee KH, Lin YF (2005) On the surface modification of microchannels for microcapillary electrophoresis chips. Electrophoresis 26:4616-4624. 137. Xiao DQ, Zhang H, Wirth M (2002) Chemical modification of the surface of poly(dimethylsiloxane) by atom-transfer radical polymerization of acrylamide. Langmuir 18:9971-9976.
160 Zhu and Wang 138. Hu SW, Ren XQ, Bachman M, Sims CE, Li GP, Allbritton N (2002) Surface modification of poly(dimethylsiloxane) microfluidic devices by ultraviolet polymer grafting. Anal Chem 74:4117-4123. 139. Hu SW, Ren XQ, Bachman M, Sims CE, Li GP, Allbritton N (2003) Crosslinked coatings for electrophoretic separations in poly(dimethylsiloxane) microchannels. Electrophoresis 24:3679-3688. 140. Hu SW, Ren XQ, Bachman M, Sims CE, Li GP, Allbritton NL (2004) Tailoring the surface properties of poly(dimethylsiloxane) microfluidic devices. Langmuir 20:5569-5574. 141. Xu F, Jabasini M, Baba Y (2002) DNA separation by microchip electrophoresis using low-viscosity hydroxypropylmethylcellulose-50 solutions enhanced by polyhydroxy compounds. Electrophoresis 23:3608-3614. 142. Henry AC, Tutt TJ, Galloway M, Davidson YY, McWhorter CS, Soper SA, McCarley RL (2000) Surface modification of poly(methyl methacrylate) used in the fabrication of microanalytical devices. Anal Chem 72:5331-5337. 143. Soper SA, Galloway M, Wabuyele M, Vaidya B, Henry A, McCarley R (2002) Surface-modification of polymer-based microfluidic devices. Analytical Chemica Acta 470:87-99. 144. Woolley AT, Sensabaugh GF, Mathies RA (1997) High-speed DNA genotyping using microfabricated capillary array electrophoresis chips. Anal Chem 69:2181-2186. 145. Simpson PC, Roach D, Woolley AT, Thorsen T, Johnston R, Sensabaugh GF, Mathies RA (1998) High-throughput genetic analysis using microfabricated 96-sample capillary array electrophoresis microplates. Proceedings of the National Academy of Sciences of the United States of America 95:22562261. 146. Shi YN, Simpson PC, Scherer JR, Wexler D, Skibola C, Smith MT, Mathies RA (1999) Radial capillary array electrophoresis microplate and scanner for high-performance nucleic acid analysis. Anal Chem 71:5354-5361. 147. Emrich CA, Tian HJ, Medintz IL, Mathies RA (2002) Microfabricated 384lane capillary array electrophoresis bioanalyzer for ultrahigh-throughput genetic analysis. Anal Chem 74:5076-5083. 148. Liu SR, Ren HJ, Gao QF, Roach DJ, Loder RT, Armstrong TM, Mao QL, Blaga I, Barker DL, Jovanovich SB (2000) Automated parallel DNA sequencing on multiple channel microchips. Proceedings of the National Academy of Sciences of the United States of America 97:5369-5374. 149. Simpson JW, Ruiz-Martinez MC, Mulhern GT Berka J, Latimer DR, Ball JA, Rothberg JM, Went GT (2000) A transmission imaging spectrograph and microfabricated channel system for DNA analysis. Electrophoresis 21:135-149. 150. Backhouse C, Caamano M, Oaks F, Nordman E, Carrillo A, Johnson B, Bay S (2000) DNA sequencing in a monolithic microchannel device. Electrophoresis 21:150-156. 151. Culbertson CT, Jacobson S C, Ramsey JM (1998) Dispersion sources for compact geometries on microchips. Anal Chem 70:3781-3789.
Genetic Analysis in Miniaturized Electrophoresis Systems 161 152. Paegel BM, Hutt LD, Simpson PC, Mathies RA (2000) Turn geometry for minimizing band broadening in microfabricated capillary electrophoresis channels. Anal Chem 72:3030-3037. 153. Paegel BM, Emrich CA, Weyemayer GJ, Scherer JR, Mathies RA (2002) High throughput DNA sequencing with a microfabricated 96-lane capillary array electrophoresis bioprocessor. Proceedings of the National Academy of Sciences of the United States of America 99:574-579. 154. Salas-Solano O, Schmalzing D, Koutny L, Buonocore S, Adourian A, Matsudaira P, Ehrlich D (2000) Optimization of high-performance DNA sequencing on short microfabricated electrophoretic devices. Anal Chem 72:3129-3137. 155. Koutny L, Schmalzing D, Salas-Solano O, El-Difrawy S, Adourian A, Buonocore S, Abbey K, McEwan P, Matsudaira P, Ehrlich D (2000) Eight hundred base sequencing in a microfabricated electrophoretic device. Anal Chem 72:3388-3391. 156. El-Difrawy SA, Lam R, Aborn JH, Novotny M, Gismondi EA, Matsudaira P, McKenna BK, O'Neil T, Streechon P, Ehrlich DJ (2005) High throughput system for DNA sequencing. Review of Scientific Instruments 76:074301. 157. Liu J, Hansen C, Quake SR (2003) Solving the "world-to-chip" interface problem with a microfluidic matrix. Anal Chem 75:4718-4723. 158. Hashimoto M, Chen PC, Mitchell MW, Nikitopoulos DE, Soper SA, Murphy MC (2004) Rapid PCR in a continuous flow device. Lab on A Chip 4:638645. 159. Kopp MU, de Mello AJ, Manz A (1998) Chemical amplification: Continuous-flow PCR on a chip. Science 280:1046-1048. 160. Chen L, Manz A, Day PJR (2007) Total nucleic acid analysis integrated on microfluidic devices. Lab on A Chip 7:1413-1423. 161. Kricka LJ, Wilding P (2003) Microchip PCR. Analytical and Bioanalytical Chemistry 377:820-825. 162. Roper MG, Easley CJ, Landers JP (2005) Advances in polymerase chain reaction on microfluidic chips. Anal Chem 77:3887-3893. 163. Zhang CS, Xing D (2007) Miniaturized PCR chips for nucleic acid amplification and analysis: latest advances and future trends. Nucleic Acids Research 35:4223-4237. 164. Woolley AT, Hadley D, Landre P, deMello AJ, Mathies RA, Northrup MA (1996) Functional integration of PCR amplification and capillary electrophoresis in a microfabricated DNA analysis device. Anal Chem 68:4081-4086. 165. Lagally ET, Emrich CA, Mathies RA (2001) Fully integrated PCR-capillary electrophoresis microsystem for DNA analysis. Lab on A Chip 1:102-107. 166. Lagally ET, Simpson PC, Mathies RA (2000) Monolithic integrated microfluidic DNA amplification and capillary electrophoresis analysis system. Sensors and Actuators B-Chemical 63:138-146. 167. Giordano BC, Ferrance J, Swedberg S, Huhmer AFR, Landers JP (2001) Polymerase chain reaction in polymeric microchips: DNA amplification in less than 240 seconds. Analytical Biochemistry 291:124-132.
162 Zhu and Wang 168. Oda RP, Strausbauch MA, Huhmer AFR, Borson N, Jurrens SR, Craighead J, Wettstein PJ, Eckloff B, Kline B, Landers JP (1998) Infrared-mediated thermocycling for ultrafast polymerase chain reaction amplification of DNA. Anal Chem 70:4361-4368. 169. Soper SA, Williams DC, Xu YC, Lassiter SJ, Zhang YL, Ford SM, Bruch RC (1998) Sanger DNA-sequencing reactions performed in a solid-phase nanoreactor directly coupled to capillary gel electrophoresis. Anal Chem 70:40364043. 170. Wang H, Chen JF, Zhu L, Shadpour H, Hupert ML, Soper SA (2006) Continuous flow thermal cycler microchip for DNA cycle sequencing. Anal Chem 78:6223-6231. 171. Breadmore MC, Wolfe KA, Arcibal IG, Leung WK, Dickson D, Giordano BC, Power ME, Ferrance JP, Feldman SH, Norris PM, Landers JP (2003) Microchip-based purification of DNA from biological samples. Anal Chem 75:1880-1886. 172. Tian HJ, Huhmer AFR, Landers JP (2000) Evaluation of silica resins for direct and efficient extraction of DNA from complex biological matrices in a miniaturized format. Analytical Biochemistry 283:175-191. 173. Wolfe KA, Breadmore MC, Ferrance JP, Power ME, Conroy JF, Norris PM, Landers JP (2002) Toward a microchip-based solid-phase extraction method for isolation of nucleic acids. Electrophoresis 23:727-733. 174. Xu YC, Vaidya B, Patel AB, Ford SM, McCarley RL, Soper SA (2003) Solid-phase reversible immobilization in microfluidic chips for the purification of dye-labeled DNA sequencing fragments. Anal Chem 75:2975-2984. 175. Deangelis MM, Wang DG, Hawkins TL (1995) Solid-phase reversible immobilization for the isolation of pcr products. Nucleic Acids Research 23:4742-4743. 176. Hawkins TL, Oconnormorin T, Roy A, Santillan C (1994) DNA purification and isolation using a solid-phase. Nucleic Acids Research 22:4543-4544. 177. Paegel BM, Yeung SHI, Mathies RA (2002) Microchip bioprocessor for integrated nanovolume sample purification and DNA sequencing. Anal Chem 74:5092-5098. 178. Manz A, Graber N, Widmer HM (1990) Miniaturized total chemical-analysis systems - a novel concept for chemical sensing. Sensors and Actuators BChemical 1:244-248. 179. Burns MA, Johnson BN, Brahmasandra SN, Handique K, Webster JR, Krishnan M, Sammarco TS, Man PM, Jones D, Heldsinger D, Mastrangelo C , Burke DT (1998) An integrated nanoliter DNA analysis device. Science 282:484-487. 180. Easley CJ, Karlinsey JM, Bienvenue JM, Legendre LA, Roper MG, Feldman SH, Hughes MA, Hewlett EL, Merkel TJ, Ferrance JP, Landers JP (2006) A fully integrated microfluidic genetic analysis system with sample-in-answerout capability. Proceedings of the National Academy of Sciences of the United States of America 103:19272-19277. 181. Bousse L, Stephane M, Minalla A, Yee H, Williams K, Dubrow R (2001) Protein Sizing on a Microchip. Anal Chem 73:207-1212.
Genetic Analysis in Miniaturized Electrophoresis Systems 163 182. CaliperLifesciences. (2002) Caliper Lifesciences 2002 Annual Report. 183. CaliperLifesciences. (2003) Caliper Lifesciences 2003 Annual Report. 13.
Chapter 5 Microfluidic Systems for Protein Separations
Anup K. Singh Biosystem Research Department, Sandia National Laboratories, Livermore, CA 94551 Correspondence should be addressed to: Anup K. Singh (
[email protected])
Keywords: Microfluidic protein separation, microfluidic liquid chromatography, multi-dimensional separation, electrophoresis
Abstract Proteins are the molecules that carry out virtually all functions in living cells. Consequently, numerous methods have been developed to separate and analyze proteins for their identification and quantification. Microfluidic devices have attracted significant attention for proteomic analysis owing to a number of advantages they offer over conventional methods including improved speed of analysis, higher resolution, increased multiplexing, and the ability to analyze minute amounts of sample.
166
Singh
5.1 Introduction Emerging research areas in biology and biotechnology, such as genomics, proteomics, and structural biology, increasingly require a large number of experiments performed in a smaller amount of time. Moreover, in most instances, these ever-increasing number of experiments need to be performed using a limiting amount of starting biological sample. These needs require a scaling down of the analysis methods. In response to this trend, analogous to the integrated circuit (IC)-chip revolution, “microfluidic chips” are starting to transform the field of biochemical analysis. While the search for “the killer application” of microfluidics continues, microfluidic chips have come a long way over the last decade and, in addition to academic research applications, they are being used in many commercial devices for analysis of DNA, RNA, proteins, and cells. In this chapter, we will focus on application of microfluidic chips for protein separations and proteomics.
5.1.1 Advantages of microfluidic chips for protein separations There are many reasons for ever-increasing applications of microchips for protein analysis as summarized below. Faster separation: Adaptation of protein analysis to a chip leads to a 10 100-fold increase in the speed of analysis. The key reason for this is the scaling down of analysis or separation dimensions compared to conventional counterparts. For example, the separation distance covered in slabgel protein electrophoresis is typically 6-10 cm. In a chip, the length of a separation channel could be less than 0.5 cm and hence, the same sample can be analyzed 12-20 times faster. Of course, this also requires a scaling down of sample size, but in many applications this is an additional advantage. For electrokinetic separations, there is another factor that improves the speed - the ability to apply higher electric fields with everything else remaining constant. In slab-gel electrophoresis, Joule heating limits the applied electric field to less than 250V/cm. In a glass chip, since glass is a good heat conductor and microchannels boast high surface area-to-volume ratios, fields as high as 750V/cm are routine, leading to another 3-5 times improvement in speed. With the chip format, improvement in speed enables analysis of a larger number of samples per unit time, as well as quicker sample-to-answer times compared to conventional methods.
Microfluidic Systems for Protein Separations
167
Smaller sample size and reagent volume: Another key advantage offered by microfluidic systems is that they require, and are capable of handling, a very small amount of sample and reagent for each analysis - a few tens or hundreds of nanoliters. These volumes are impossible to analyze in conventional microtiter plates or vials as they will evaporate within seconds. This advantage also results from the scaled-down dimensions offered in a microchip with sealed microchannels. In addition, the small size of the chips themselves, make it possible to reduce the footprint of the analysis device. It is anticipated that in the not too distant future, we will have a personal or portable protein analyzer based on a protein analysis chip.
5.1.2 Limitations of microfluidic chips in proteomics applications The key limitations of microchips are side-effects of the advantages they offer. For example, while miniaturization enables faster analysis, it reduces the peak-capacity of chip-based separations. Current microchips cannot perform the equivalent of a conventional 2-dimensional separation on a complex sample such as cell lysate or blood serum. For similar reasons, chips also cannot handle complex dirty samples. Limited peak capacity makes it hard to resolve peaks if there are too many components present or if one component is present in huge excess. Another factor, resulting from the small cross-sections involved, is that it is relatively easy to clog the channels with particulates or aggregates in the sample. It is also very easy to foul the separation channels, which reduces run-to-run reproducibility and limits reuse of the chip. Scaling down of dimensions results in a smaller path-length for optical detection of analytes as well. So while limit of detection (LOD) in terms of amounts (or moles) is superb in chips, the concentration sensitivity can be very poor. This makes detection methods such as absorbance unsuitable for microfluidic analysis. One typically has to use fluorescence, chemiluminescence or electrochemical detection methods with chips.
5.1.3 Substrates used for proteomic analysis Glass, fused-silica, or quartz have been the most widely used substrates and have been commercialized by companies such as Agilent, Caliper and Bio-Rad. Other vendors for these chips are Micronit and Micralyne. Plastic
168
Singh
chips have also been developed - owing to their inexpensiveness and simple fabrication. The key criterion used for suitable polymers to fabricate a chip is the requirement that they are transparent in spectral regions of interest for available detectors. Hence, the possible choices are acrylates such as Polymethyl methacrylate (PMMA), Polydimethylsiloxane (PDMS), cyclic olefin copolymer (COC), and polycarbonate (PC). The most widely-used polymeric substrates are PMMA and PDMS. In the early days, researchers had to find fabrication facilities in-house or through a collaborator who had access to one. This really limited the places where microfluidic research could be done. Now, with wider acceptance of chip-based analysis, there are a number of commercial vendors that provide standard as well as custom chips for virtually any application.
5.2. Microfluidic Chips for Protein Separation Most of the early applications of chips for proteomic applications utilized open-channel separations. The first successful adaptation of a protein analysis technique to the microchip was capillary zone electrophoresis (CZE) because it is perhaps the simplest and easiest method to adapt to a chip. Other protein separation applications adapted to microfluidic chips include: capillary gel electrophoresis (CGE), Sodium dodecyl sulfate (SDS) polyacrylamide gel electrophoresis (PAGE), isoelectric focusing (IEF), micellar electrokinetic chromatography (MEKC), isotachophoresis (ITP) and open-tubular chromatography. Figure 5.1 lists the protein separation techniques that have been adapted to chips.
Microfluidic Systems for Protein Separations
Capillary Electrophoresis Gel-Based: CGE PAGE & SDS-PAGE Open-channel: CZE ITP IEF MEKC
169
Liquid Chromatography Ion Exchange Size Exclusion Reversed-Phase Electrochromatography Hydrophobic Interaction Affinity
Multi-Dimensional Techniques LC-MS CE-MS LC-CE 2D Gel Electrophoresis
Fig. 5.1 Protein separation techniques adapted to microfluidic chips.
5.2.1 Microchip-based electrophoretic techniques 5.2.1.1.Capillary zone electrophoresis of proteins
Capillary zone electrophoresis (CZE) separates analytes based on their charge/mass ratio and is carried out in a capillary filled with an aqueous buffer. CZE offers extremely high resolution and hence, has been widely used for the separation of proteins and peptides. CZE has been adapted to a planar microchannel format by a number of researchers [1-10]. Chip-based CZE allows better control of sample introduction and leads to better performance in terms of speed and efficiency over the conventional capillarybased separation. Microchip-based CZE has also been multiplexed and integrated with other forms of separation to enable multidimensional separations. Chips for CZE of proteins, similar to capillary-based separations, require that channel walls be coated covalently or dynamically to reduce non-specific adsorption of analytes. Proteins can adsorb non-specifically on glass surfaces due to electrostatic forces (for cationic proteins or even anionic or neutral ones with cationic domains) or hydrophobic interactions. Fortunately, many different covalent chemistries or dynamic coating reagents, e.g., polyethylene glycol (PEG) based coating, have been developed for fused-silica capillaries that can readily be used for glass, fused-
170
Singh
silica or quartz chips [11]. However, for plastic chips, either hydrophilic substrates need to be used or special coatings developed to minimize adsorption [12]. 5.2.1.2. Sizing of proteins
SDS-PAGE is one of the most commonly used techniques for analysis of proteins. It separates proteins based on their size and is typically performed in slab-gels. It is used to determine protein purity, molecular weight and also to identify proteins after a processing step such as affinity separation. The current slab-gel devices require manual loading of samples using pipettes and a typical run is completed in 45 min to an hour. After SDS-PAGE, the gel is stained with a dye, destained to remove excess dye and then imaged in a scanner. These steps take an additional few hours to complete. SDS-PAGE has not substantially changed in the 30 years since its introduction. Adaptation of protein sizing to microchips has overcome many of the limitations of slab-gel electrophoresis by making the separations faster. The most popular chip-based sizing technique is gel electrophoresis using a liquid sieving gel where the gel is made up of linear polyacrylamide or another hydrophilic polymer such as polyethylene glycol [13 - 14]. Unlike slab-gel and CGE in capillaries, the entire process of protein loading, SDSPAGE and detection is integrated and automated. Furthermore, the entire analysis could be completed in less than a minute per sample. This is also an application that as been commercialized as perhaps the most promising microfluidic-chip based device. Three companies, Caliper, Agilent and Bio-Rad, manufacture instruments relying on Caliper’s protein and DNA sizing chips. Slab gel based SDS-PAGE has also been adapted to the microchip format. Han and Singh [15] describe a photopatterning technique to cast in situ crosslinked polyacrylamide gel in a microchannel to perform SDS-PAGE. A fluorescent protein marker sample was separated in less than 30 sec in less than 2 mm of channel length. UV-patterned polyacrylamide gel provides higher sieving power and sample stacking effect, therefore yielding faster separation in a chip. The use of solid polyacrylamide gel instead of liquid sieving matrix in SDS-PAGE also enables easier integration with another separation module.
Microfluidic Systems for Protein Separations
171
5.2.1.3. IEF of proteins
Over the past 30 years, isoelectric focusing (IEF) has grown to be a prominent method of separation and detection in proteome research. The technique separates and focuses analytical species along an established pH gradient based on their different isoelectric points (pI), the pH at which an analyte’s net charge is zero. IEF is commonly used in multi-dimensional separations (i.e. IEF-PAGE) as it provides an orthogonal separation mechanism to size-based fractionation. The earliest microscale adaptation of IEF was done by the Pawliszyn group using a very simple chip containing one channel with an inlet and outlet [16]. The sample was mixed with ampholytes - soluble amphoteric compounds that establish a pH gradient upon application of an electric field - and then loaded into the chip. Voltage was applied to carry out IEF and the entire chip was imaged for detection of the focused bands. Since that initial demonstration, many other articles have also appeared on microscale IEF [15,17-20]. Recently, IEF was achieved on a microchip containing an immobilized pH gradient (IPG), enabling improved pH gradient stability over conventional free-flow methods [21]. The authors established the pH gradient via diffusion of pKspecific acrylamido buffers along a flow-restricted channel. Precise control over boundary conditions and the resulting gradient is achieved by continuous flow of stock solutions through side channels flanking the gradient segment (Figure 5.2). Once the desired gradient is established, it is immobilized via photopolymerization. Rapid (< 20 minutes) isoelectric focusing of several fluorescent pI markers and proteins is demonstrated across pH 3.8 – 7.0 in µIPG’s without the addition of carrier ampholytes, using both denaturing and non-denaturing conditions.
172
Singh
(a) Low pH
High pH Polymerized Membrane
100µm Immobiline Diffusion
6mm
(b) (i)
4.0
4.5
(ii)
(iii)
5.1
Tfer
BSA
500µm
OVA
5.5
6.2
6.6 6.8
CA PhB Hb
GFP
Fig. 5.2 a)µ IPG fabrication schematics. Gravity was used to provide continuous pressure driven flow from filled reservoirs through the side channels of the glass microchip. A thin polyacrylamide membrane prohibits flow through the 6mm separation channel yet allows equilibration of Immobiline species via diffusion. Entire chip is photopolymerized with UV light following equilibration. b) Composite fluorescence IEF images spanning length of 6mm-long µIPG. (i) Fluorescent pI markers. (ii) AF 647-labeled proteins: Bovine serum albumin (BSA), Transferrin (Tfer), Carbonic anhydrase (CA, bovine), Phosphorylase B (PhB, rabbit muscle), Hemoglobin (Hb, bovine). (iii) AF 488-labeled ovalbumin (OVA) and green fluorescent protein (GFP). Reprinted with permission from [21]. Copyright (2008) American Chemical Society.
5.2.2 Microchip chromatography Chromatography, e.g., high performance liquid chromatography (HPLC), because of its outstanding separation power and versatility, is one of the most common analysis methods for proteins and peptides. There are many
Microfluidic Systems for Protein Separations
173
types of chromatography available, as shown in Figure 5.1, enabled by the wide variety of chromatographic media available. While HPLC studies to date have been implemented in a macroscopic format (in columns with typical capacities and internal diameters of 1 mL and 4.6 mm, respectively) miniaturization presents the potential for several advantages. Improvements have been realized by the advent of microbore and capillarybased HPLC columns, with internal diameters of 100 µm to 2 mm, although injection and detection methods have remained largely unchanged. Significant efforts have been directed towards adapting chromatographic techniques to the microchip format to allow reduction of injection size, a critical advantage when samples are expensive, difficult to generate, or when the scientific question requires minimization of the volume (e.g., single-cell analysis). Reduction of column dimensions also reduces the system flow rate, which leads to improved signal-to-noise when chromatographic separations are connected to concentration-sensitive detectors such as electrospray injection mass spectrometry [22]. A major roadblock in adaptation of chromatography to microchips has been the issue of placement of chromatographic media in the channels. Uniform packing of channels with particles is irreproducible, hard to implement and requires fabrication of shelves and side channels. In the early days of adaptation of chromatography to chips, methods were proposed that obviate the need for packing channels by using open-channel separations. The two examples are- micellar electrokinetic chromatography (MEKC) and open-tubular chromatography. In MEKC, SDS micelles are used as a psuedostationary phase for partitioning of analytes based on their hydrophobicity [23]. While MEKC provides reasonable separation resolution, it does not match the flexibility, reproducibility, peak capacity and resolution of packed-bed reversed-phase chromatography. The biggest drawback of MEKC is that it does not allow the equivalent of gradient elution separations which are so widely used in the resolution of complex peptide and protein mixtures. Open tubular columns with the stationary phase supported on the channel walls is another way to circumvent the use of particles and the accompanying packing problems. Although open tubular columns were first described nearly three decades ago, [24] they have never become popular in liquid chromatography. This is probably because channel widths of 2 µm or less are required to deal with the limited rates of mobile-phase mass transfer in liquid chromatography [25]. Columns with this small diameter are also easily plugged, the loading capacity is extremely small, gradient elution with positive displacement pumps is difficult, and the optical path length used for detection is very short.
174
Singh
Successful adaptation of packed-bed chromatography to chips was enabled by formation of UV-initiated porous polymers in microchips pioneered by [26] and used by several other research groups [27-28]. In situ casting of polymer monoliths in microchannels reproducibly affords uniform packed beds, therefore eliminating the difficulties associated with packing silica beads and the need for retaining frits. The availability of a wide range of monomers enables critical stationary phase properties such as charge and hydrophobicity to be easily tuned to meet the specific demands of separating many types of analytes. Photopolymerization enables the patterning of a chromatographic media in the microchip, analogous to photolithography, using a mask for optimal design of injection, separation and detection manifolds. These monoliths can be cast in situ in less than 10 minutes, are robust and reproducible with respect to separation characteristics. The microchip in Figure 5.3 was used for analysis of bioactive peptides and amino acids and resulted in separations that were fast (6 peptides in 45 sec), efficient (up to 600,000 plates/m) and reproducible (run-to-run variability <3%) [27]. 4
b)
Kr-ion Laser (413 nm)
Fluorescence
a)
3 6
5
1 2
0
10
20
30
40
50
Seconds
Fig. 5.3 A microchip with photopolymerized monoliths for reversed-phase electrochromatography of peptides. a) The schematic of the chip containing isotropically-etched channels that are 20-40 micron deep, ~100 micron wide, and 5 cm long. The inset shows scanning electron micrograph of porous polymer monolith cast in the channels by photoinititaion using a UV-lamp. b) Separation of 6 bioactive peptides in the microchip in less than 45 seconds. Peaks are: (1) papain inhibitor, (2) proctolin, (3) Casein fragment 90-95, (4) Ile-angiotensin III, (5) angiotensin III, (6) Gly-Gly-Gly. Adpated with permission from [27]. Copyright (2002) American Chemical Society.
Microfluidic Systems for Protein Separations
175
5.3 Integrated Analysis in Microchips The applications discussed earlier have focused mainly on implementation of single separation techniques in a microchip. However, analogous to an IC chip, the power of microfluidic chips lies in its ability to integrate multiple analysis steps in one chip or perform many parallel analyses simultaneously. For example, chips have been developed to perform multiple immunoassays in a single chip. Similarly, chips have been developed to integrate sample treatment with analysis, [29-30] an important step towards developing sample-to-answer chips. Many biological samples are too complex (e.g., blood, cell lysate) to be fully analyzed using just one separation technique. In these cases, one resorts to integrating two or more orthogonal separation techniques for analyzing a sample. Key examples are two-dimensional liquid chromatography and two-dimensional gel electrophoresis. In the sections below, we provide a discussion of some of the “integrated” analyses implemented in microchips.
5.3.1 Integration of sample preparation with analysis While many protein analysis methods such as separations, immunoassays, etc. have been successfully adapted to microchips, in most cases sample pretreatment is performed off-chip. Integration of sample preparation with analysis serves many purposes including easier automation, improvement in speed of analysis, and reduction in sample loss. Numerous approaches have been developed to incorporate functions such as sample cleanup, sample concentration, mixing, and reaction prior to analysis in microchips over the past decade [31-33]. Review articles describing sample pretreatment methods in microfluidic systems have been recently published [3436]. Below we describe an example of integrated sample preparation and sample concentration prior to separation. There are a number of reasons why sample concentration prior to analysis is a crucial step in the development of multi-functional integrated microfluidic devices. First and foremost is that preconcentration of sample enables detection of trace or low-abundant species. This is of particular importance in many fields including clinical diagnostics, proteomics, forensics, environmental monitoring, and biodefense applications. A second motivation for preconcentration arises from the fact that micrometer dimensions of the fluidic channels lead to poorer sensitivities for optical detection than their conventional scale counterparts. Preconcentration not
176
Singh
only improves detection sensitivity but also improves the reliability of analysis by significantly increasing signal-to-noise ratios. In the previous sections, we discussed implementation of protein sizing in protein chips. To improve the sensitivity of SDS-PAGE, Hatch et al. [29] reported on a novel integrated preconcentrator that allows 1000-fold improvement in detection limit enabling detection of proteins at ~50 fM concentrations. This was achieved by trapping proteins on a size-exclusion membrane that was fabricated in the sample loading path of a chip. The membrane had a MW cut-off of 10 kDa and hence when the sample was electrophoresed through the membrane, proteins in the sample were trapped on the upstream side of the membrane. Later the concentrated proteins were eluted, by reversing the electric field, and separated in a channel containing the separation gel. Figure 5.4 shows the schematic of concentration, elution and separation of protein mixtures. The extent of protein preconcentration is easily tuned by varying the voltage during injection or by controlling the sample volume loaded. The integrated preconcentrationsizing approach facilitates analysis of low-abundant proteins that cannot be otherwise detected and the approach has been extended to lowering of detection limits for other applications such as clinical diagnostics [37]. A similar approach has also been used by other researchers. Khandurina et al. demonstrated a size-exclusion approach for concentrating DNA [38] and more recently for concentrating proteins [39], wherein a silicate membrane was deposited between the glass cover plate and silicon substrate of a microchip. Charged molecules are trapped on this nanoporous membrane prior to separation by CZE.
Microfluidic Systems for Protein Separations
177
D
Fig. 5.4 The sequential process of protein preconcentration (A), elution (B), and separation (C). The fluorescence micrographs show the distribution of labeled protein (visible only after preconcentration) at different time-points of the integrated process. (A) During the preconcentration step, an electric field drives the transport of protein SDS complexes toward the size-exclusion membrane where they become trapped and accumulate as long as the field is applied. The fluorescence micrographs show labeled BSA accumulating at the membrane (t0 = 0s, t1 = 30s, and t2 = 120s preconcentration time). (B) Following the preconcentration step, the field across the membrane is reversed, redirecting proteins away from the membrane (t3 = 120.5s, t4 = 121s). (C) During SDS-PAGE, proteins migrate into the separation channel where they are size-separated. (D) Electropherograms for a concentrated mixture of 4 proteins separated by SDS-PAGE. Without preconcentration, the proteins were just above the detector threshold (control). Adapted with permission from [29]. Copyright (2006) American Chemical Society.
5.3.2 Multi-dimensional separation in microchips The monitoring of protein expression profiles in a biological fluid such as a cell lysate or serum remains a very challenging task because of the large number of proteins (>100,000 in human proteome); the wide dynamic range of these proteins; and the variability of gene products (splicing variants, N- and C-terminal truncations, co- and post-translational modifications, etc.). No single separation method has the ability to resolve these
178
Singh
large numbers (>1000) of components and hence, many multi-dimensional techniques have been developed. The predominant approach in conventional scale experiments has been separation by two-dimensional gel electrophoresis (2-D gel), followed by identification of protein spots using sensitive mass spectrometry techniques. An alternative, and progressively more popular, approach to 2-D gel includes a comprehensive chromatographic separation of the proteolytic fragments derived from intact proteins, followed by mass spectral identification and data base searching. This can be achieved using a two-dimensional liquid chromatography approach whereby peptides are fractionated on a strong cation exchange column, followed by an extended gradient elution on a C18 reverse phase column. Microfluidic systems hold a great promise for realizing multidimensional separations in a single integrated system. To this end, several groups have explored the application of microfluidics to multidimensional peptide and protein separations. For example, Rocklin and co-workers [40] have demonstrated 2-D separation of peptide mixtures in a microfluidic device using MEKC and CZE as the first and second dimensions, respectively. While MEKC and CZE are not perfectly orthogonal, they are perhaps the easiest to implement in a microchip with open channels and, when combined, still provide high peak capacity. In this example, a mixture of peptides is first separated by MEKC and then the effluent is sampled every 4 sec with a 0.3 sec injection to perform CZE in the second dimension, with total analysis times of less than 10 min for tryptic peptides. Approximately 10% of the sample injected into the first dimension is sampled by the second dimension and the peak capacity of the two-dimensional separations was estimated to be in the 500-1000 range. In a similar approach, Gottschlich et al. [41] reported on combining open-tubular reversed-phase chromatography with CZE for two-dimensional separation of peptides. A number of groups have also succeeded, at least partially, in adapting variants of 2D gel electrophoresis to microchips. True miniaturization of 2D gel electrophoresis has not been achieved yet because of the difficulties involved in the transfer of analytes from the first dimension to the second. This results mostly because of the incompatibility of the buffer systems used in the two separations - IEF needs highly-controlled mixture of ampholytes with low ionic conductivity while SDS-PAGE uses a highly conductive buffer. Leakage of SDS into IEF channel can completely ruin the IEF separation. Herr et al. [42] first demonstrated the integration between liquid-phase IEF and free solution CE by sequentially transferring samples
Microfluidic Systems for Protein Separations
179
from the IEF channel to an orthogonal CE channel. Once transferred to the CE channel, samples can be further separated based on their electrophoretic mobilities. Meanwhile, Wang et al. [43] implemented softlithography microfluidic valves to couple carrier ampholyte IEF with either CGE or CZE in a PDMS microchip. By active microvalve control, IEF focused samples can be isolated and selectively transferred to the second dimension separation channel. Further success in this area includes a plastic microfluidic network developed by Li et al. [44] and a multichannel differential gel electrophoresis (DIGE) platform presented by Emrich et al. [45]. Li et al. coupled nonnative IEF with SDS gel electrophoresis in a multi-channel platform. By controlling the channel resistance, sieving gel solution was first loaded into multiple separation channels. To increase the peak capacity (~1700), SDS solution was introduced between the first dimension IEF and second dimension gel electrophoresis to form SDS complexes [44]. Emrich et al. described a microfluidic platform integrating liquid-phase IEF and the second dimension CE with a fluid barrier created by passive valve structures (shallow etches). Cell lysate from 440,000 E. coli cells were successfully analyzed by the micro-DIGE analyzer as shown in Figure 5.5 [45].
Fig. 5.5 Schematic of the microchip to perform two-dimensional gel electrophoresis. The chip comprises an arced, 3.75 cm long horizontal channel for first dimension isoelectric focusing (IEF) that is punctuated with 20 6.8 cm long vertical channels through which focused proteins are separated in the second dimension by native gel electrophoresis. Reprinted with permission from [45]. Copyright (2007) American Chemical Society.
180
Singh
5.3.3 Chips integrated with mass spectrometry Recent advances in speed, sensitivity and throughput of mass spectrometers has made them the detector of choice for many proteomic applications. The mass spectrometer as a detector offers many advantages including label-free detection and identification of a large number of peptides (and proteins) in a single run. Because of its high peak capacity, multidimensional liquid chromatography interfaced with mass spectromtery has become the technique of choice for genome-wide identification of proteins. Integration of microfluidic chip-based separations (e.g, chromatography) with mass spectrometry is attractive for a number of reasons including 1) ability to analyze smaller samples, 2) improved speed of analysis, and 3) facile integration of steps such as enrichment, separation and electrospray. The microchip interface to mass spectrometry has been achieved by a number of methods including creating a guided channel to the nanoelectrospray emitter using butted capillary [46], the double-etching procedure [47], polymer casting [48], and microdrilling [49-50]. One commercial product that allows integration of chip to a mass spectrometer is the Agilent's HPLC-Chip technology (www.agilent.com). The HPLC-Chip integrates sample preparation, separation, and electrospray tip on a single chip. The reusable chip is machined out of polyimide and contains a channel that serves as both separation column and electrospray nozzle. On-chip integration of sample, enrichment, and electrospray significantly reduces the number of fittings, connections, valves, and tubing required for conventional HPLC-MS and leads to higher-resolution separation and a more sensitive detection of peptides and other analytes.
5.4. Future Directions Microfluidic chips have made a significant impact in the field of protein analysis research and will continue to be used because of the numerous advantages they offer namely, faster, better, and cheaper analysis. Many conventional work-horse techniques such as zone electrophoresis, gel electrophoresis, isoelectric focusing, and limited forms of chromatography have been adapted to microchip format and the list will continue to grow. The biggest power of microfluidic chips, the ability to perform integrated and parallel operations, has not been realized yet. It is anticipated that in the future, progressively the focus will be on developing chips that integrate multiple functions and go towards realizing “sample-to-answer” de-
Microfluidic Systems for Protein Separations
181
vices. To date, focus has been on adapting single analysis techniques to a chip. In future, emphasis will shift to multi-dimensional analysis of proteins in a single chip and interfacing those separations with detectors such as mass spectrometers. While some successes have been achieved in integrating microfluidic chips to mass spectrometers, a lot of work needs to be done to achieve interfaces that are reproducible and easy to use. One of the problems has been that current mass spectrometers do not have the sensitivity to analyze the minute amounts that are typically analyzed in chips. With the development of more sensitive mass spectrometers and concomitant improvement in chip-based protein analysis and interfaces, it should be possible to have microfluidic devices that can be connected to any mass spectrometer.
References 1. Effenhauser C S (1998) in Microsystem Technology in Chemistry and Life Sciences, eds. Manz, A. & Becker, H. (Springer, Heidelberg) p51-82. 2. Effenhauser CS, Bruin GJM, Paulus A (1997) Integrated chip-based capillary electrophoresis. Electrophoresis 18:2203-2213. 3. Van den Berg A, Bergveld P eds. (1995) Micro Total Analysis Systems (Kluwer, Boston). 4. Harrison JD, Fluri K, Seiler K, Fan Z, Effenhauser CS Manz A (1993) Micromachining a Miniaturized Capillary Electrophoresis-Based Chemical Analysis System on a Chip. Science 261:895-897. 6. Boone TD, Hooper HH, eds. (1998) Micro Total Analysis Systems `98 (Kluwer, Boston), p. 257-260. 7. Duffy, D. C., McDonald, J. C., Schueller, O. J. & Whitesides, G. M. (1998) Rapid Prototyping of Microfluidic Systems in Poly(dimethylsiloxane). Anal Chem 70:4974-4984. 8. Manz A, Harrison JD, Verpoorte E, Widmer MH (1993) Planar Chips Technology for Miniaturization of Separation Systems: A developing perspective in chemical monitoring. Adv Chromatogr 33:1-65. 9. Chiem N, Harrison JD (1997) Microchip-Based Capillary Electrophoresis for Immunoassays: Analysis of Monoclonal Antibodies and Theophylline. Anal Chem 69:373-378. 10. Colyer CL, Mangru SD, Harrison DJ (1997) Microchip-based capillary electrophoresis of human serum proteins. J Chromatogr A 781:271-276. 11. Hjertén S (1985) High-performance electrophoresis : Elimination of electroendosmosis and solute adsorption. J Chromatogr 347:191-198. 12. Liu J, Lee ML (2006) Permanent surface modification of polymeric capillary electrophoresis microchips for protein and peptide analysis. Electrophoresis 27:3533-46.
182
Singh
13. Yao S, Anex DS, Caldwell WB, Arnold DW, Smith KB, Schultz PG (1999) SDS capillary gel electrophoresis of proteins in microfabricated channels. Proc Natl Acad Sci USA 96:5372-5377. 14. Bousse L, Mouradian S, Minalla A, Yee H, Williams K, Dubrow R (2001) Protein Sizing on a Microchip. Anal Chem 73:1207-1212. 15. Han J, Singh AK (2004) Rapid protein separations in ultra-short microchannels: microchip sodium dodecyl sulfate-polyacrylamide gel electrophoresis and isoelectric focusing. J Chromatogr A 1049:205-9. 16. Wu J, Pawliszyn J (1992) Application of capillary isoelectric focusing with universal concentration gradient detector to the analysis of protein samples. J Chromatogr 608:121-30. 17. Herr AE, Molho JI, Drouvalakis KA, Mikkelsen JC, Utz PJ, Santiago JG, Kenny TW (2003) On-chip coupling of isoelectric focusing and free solution electrophoresis for multidimensional separations. Anal Chem 75:1180-7. 18. Tan W, Fan ZH, Qiu CX, Ricco AJ, Gibbons I (2003) Miniaturized capillary isoelectric focusing in plastic microfluidic devices. Electrophoresis 23:363845. 19. Li Y, Buch JS, Rosenberger F, DeVoe DL, Lee CS (2004) Integration of isoelectric focusing with parallel sodium dodecyl sulfate gel electrophoresis for multidimensional protein separations in a plastic microfluidic network. Anal Chem 76:742-8. 20. Xu Y, Zhang CX, Janasek D, Manz A (2003) Sub-second isoelectric focusing in free flow using a microfluidic device. Lab Chip 3:224-7. 21. Sommer GJ, Singh AK, Hatch AV (2008) On-Chip Isoelectric Focusing Using Photopolymerized Immobilized pH Gradients. Anal Chem 80:3327-33. 22. Reichmuth DS, Shepodd TJ, Kirby BJ (2005) Microchip HPLC of Peptides and Proteins Anal Chem 77:2997–3000. 23. Terabe S, Otsuka K, Ichikawa K, Tsuchiya A, Ando T (1984) Electrokinetic separations with micellar solutions and open-tubular capillaries. Trends Anal Chem 56:111-113. 24. Nota G, Marino G, Buonocore V, Ballio A (1970) Liquid-solid Chromatography with Open Glass Capillary columns. J Chromatogr 46:103. 25. Knox JH, Gilbert MT (1979) Kinetic Optimisation of Straight Open-tubular Liquid Chromatography. J Chromatogr 186:405-418. 26. Rohr T, Yu C, Davey MH, Svec F, Fréchet JM (2001) Porous polymer monoliths: simple and efficient mixers prepared by direct polymerization in the channels of microfluidic chips. Electrophoresis 22:3959-67. 27. Throckmorton DJ, Shepodd TJ, Singh AK (2002) Electrochromatography in microchips: reversed-phase separation of peptides and amino acids using photopatterned rigid polymer monoliths. Anal Chem 74:784-9. 28. Bedair M, Oleschuk RD (2006) Lectin affinity chromatography using porous polymer monolith assisted nanoelectrospray MS/MS. Analyst 131:1316-21. 29. Hatch AV, Herr AE, Throckmorton DJ, Brennan JS, Singh AK (2006) Integrated preconcentration SDS-PAGE of proteins in microchips using photopatterned cross-linked polyacrylamide gels. Anal Chem 78:4976-84.
Microfluidic Systems for Protein Separations
183
30. Waters LC, Jacobson SC, Kroutchinina N, Khandurina J, Foote RS, Ramsey JM. Microchip device for cell lysis, multiplex PCR amplification, and electrophoretic sizing. Anal Chem 1998b;70:158-62. 31. Auroux PA, Iossifidis D, Reyes DR, Manz A (2002) Micro Total Analysis Systems. 2. Analytical Standard Operations and Applications. Analytical Chemistry 74:2637-2652. 32. Reyes DR, Iossifidis D, Auroux PA, Manz A (2002) Micro Total Analysis Systems. 1. Introduction, Theory, and Technology. Analytical Chemistry 74:2623-2636. 33. Vilkner T, Janasek D, Manz A (2004) Micro Total Analysis Systems. Recent Developments. Analytical Chemistry 76:3373-3386. 34. Song S, Singh AK (2006) On-chip sample preconcentration for integrated microfluidic analysis. Analytical and Bioanalytical Chemistry 384:41-43. 35. Lichtenberg J, deRooij NF, Verpoorte E (2002) Sample pretreatment on microfabricated devices. Talanta 56:233-266. 36. de Mello AJ, Beard N (2003) Focus. Dealing with real samples: sample pretreatment in microfluidic systems. Lab on a chip 3:11N-19N. 37. Herr AE, Hatch AV, Throckmorton DJ, Tran HM, Brennan JS, Giannobile WV, Singh AK (2007) Microfluidic immunoassays as rapid saliva-based clinical diagnostics. Proc Natl Acad Sci USA 104:5268-73. 38. Khandurina J, Jacobson SC, Waters LC, Foote RS, Ramsey JM (1999) Microfabricated Porous Membrane Structure for Sample Concentration and Electrophoretic Analysis. Anal Chem 71:1815-1819. 39. Foote RS, Khandurina J, Jacobson SC, Ramsey JM (2005) Preconcentration of proteins on microfluidic devices using porous silica membranes. Anal Chem 77:57-63. 40. Rocklin RD, Ramsey RS, Ramsey JM (2000) A microfabricated fluidic device for performing two-dimensional liquid-phase separations. Anal Chem 72:5244-9. 41. Gottschlich N, Jacobson SC, Culbertson CT, Ramsey JM (2001) Twodimensional electrochromatography/capillary electrophoresis on a microchip. Anal Chem 73:2669-74. 42. Herr AE, Molho JI, Drouvalakis KA, Mikkelsen JC, Utz PJ, Santiago JG, Kenny TW (2003) On-chip coupling of isoelectric focusing and free solution electrophoresis for multidimensional separations. Anal Chem 75:1180-7. 43. Wang YC, Choi MH, Han J (2004) Two-dimensional protein separation with advanced sample and buffer isolation using microfluidic valves. Anal Chem 76:4426-31. 44. Li Y, Buch JS, Rosenberger F, DeVoe DL, Lee CS (2004) Integration of isoelectric focusing with parallel sodium dodecyl sulfate gel electrophoresis for multidimensional protein separations in a plastic microfluidic network. Anal Chem 76:742-8. 45. Emrich CA, Medintz IL, Chu WK, Mathies RA (2007) Microfabricated twodimensional electrophoresis device for differential protein expression profiling. Anal Chem 79:7360-6.
184
Singh
46. Figeys D, Aebersold R (1998) Nanoflow solvent gradient delivery from a microfabricated device for protein identifications by electrospray ionization mass spectrometry. Anal Chem 70:3721–3727. 47. Zhang B, Liu H, Karger BL, Foret F (1999) Microfabricated devices for capillary electrophoresis-electrospray mass spectrometry. Anal Chem 71:3258– 3264. 48. Liu H, Felten C, Xue Q, Zhang B, Jedrzejewski P, Karger BL, Foret F (2000) Development of multichannel devices with an array of electrospray tips for high-throughput mass spectrometry. Anal Chem 72:3303–3310. 49. Bings NH, Wang C, Skinner CD, Colyer CL, Thibault P, Harrison DJ (1999) Microfluidic devices connected to glass capillaries with minimal dead volume. Anal Chem 71:3292–3296. 50. Li J, Thibault P, Bings NH, Skinner CD, Wang C, Colyer CL, Harrison DJ (1999) Integration of microfabricated devices to capillary electrophoresiselectrospray mass spectrometry using a low dead volume connection; application to rapid analyses of proteolytic digests. Anal Chem 71:3036–3045.
Chapter 6 Microfluidic Systems for Cellular Applications
H. Tavana, Y.-K. Chung, C.-H. Kuo and S. Takayama Department of Biomedical Engineering, University of Michigan, Ann Arbor, MI 48109 Correspondence should be addressed to: Prof. S. Takayama (
[email protected])
Keywords: Microfluidics, Cell Isolation, Cell Culture, Cell Analysis, Cellular Microenvironment, Cell Signaling
Abstract Cells are the basic building units of the vast diversity of living organisms. Study of cells and their functions is of great importance in pure and applied sciences and spans a wide range of areas including cell biology, human physiology, and tissue engineering. Much of our knowledge of intricate cellular behavior and functions is due to various techniques that allow in vitro manipulation of cells. Handling of cells outside the body often involves the three main steps of isolation, culture, and analysis. Existing traditional methods have facilitated these processes and revealed a wealth of information. Nevertheless, recent advances in our understanding of various aspects of cellular systems point to the inadequacy of conventional techniques. For example, conventional cell culture systems do not mimic intricate in vivo cellular microenvironments and as a result, many cellular phenotypes are lost. The emergence of microfluidic technology in the past
186
Tavana, Chung, Kuo, and Takayama
decade and its compatibility with all three cell handling processes has given a new spin to study of cells in various contexts. Microfluidic settings for the isolation, culture, and analysis of cells offer advantages over their conventional counterparts by providing a more physiologic microenvironment for cells, enabling various biologically important culture conditions (e.g. static and dynamic), and being economically sounder. The commercialization of microfluidic cell-based platforms also shows a promising trend and new products steadily find their way into the market.
6.1 Introduction Cells are the simplest structural units into which a complex multicellular organism can be divided and still retain the functions characteristic of life. Cells of the body arrange themselves in various combinations to form a hierarchy of organized structures ranging from tissues, to organs, to organ systems, and finally to a total organism. The response of an organism to disease, injury, or therapy is therefore a collective response of its cells. For decades, in vitro cell culture has served as a gateway to explore various aspects of cellular biology. The ease of introducing systemic alterations to the environment of cells and the possibility of modulating the phenotype of cells has prompted an unprecedented expansion of in vitro systems with the aim of designing cell-based assays, sensors, and substituents of native tissues. Engineered in vitro constructs have also served as models for quantitative studies of cell and tissue response to genetic alterations, drugs, hypoxia, and mechanical stimuli. Studies of cells often involve three key steps: separation of a cell type of interest from a heterogeneous population, culture of the isolated cells, and performing biochemical analyses on them. Standard procedures exist for each of these steps; however, they entail a number of limitations that are briefly discussed below. Conventional methods for separation of a cell type of interest from a heterogeneous population are based on differences in physicochemical properties of different cell types including size, density, and electric charge [14]. However, these properties vary only slightly from one cell type to another, making cell separation processes inefficient. Immunologic techniques, such as magnetic cell sorting and fluorescence-based flow cytometry, involve selectively attaching magnetic or fluorescent particles to cells
Microfluidic Systems for Cellular Applications
187
and yield higher efficiencies [1,5]. Nevertheless, these separation methods require preparatory and incubation time periods, actuation and detection systems, and post-processing of cells such as the removal of labeling. Overall, traditional approaches for cell separation are costly, timeconsuming, mechanically complex, and require trained personnel for operation and maintenance. They can also be lethally or sub-lethally damaging to cells. The history of cell culture dates back to early twentieth century. The original drive for the development of cell cultures was to investigate various physiological events such as nerve development [6]. It was in the 1940s and 1950s that animal cell culture became a routine laboratory technique. Historically, in vitro culture of animal cells has been conducted on simple, homogeneous surfaces such as those of Petri dishes. It is well known that the cells in these in vitro environments often do not express the properties characteristic of the same cell type in vivo. The problem stems from the lack of physiologic cellular microenvironment in these in vitro culture systems. Inside the body, cells reside in a tissue specific microenvironment and are subject to multiple cues that vary in time and space, including gradients of various soluble factors (hormones, growth factors, nutrients, and inorganic ions), biochemical and mechanical interactions with the extracellular matrix (ECM), and direct cell-cell contacts. Conventional culture systems do not provide such a physiologic environment for cells and as a result, only limited information can be gained about the phenotype of cells grown in these systems. The following example elucidates this point. In a cell culture dish, the volume of culture media above cells is significantly greater than the volume taken up by the monolayer of cultured cells. Continuous secretion of signal molecules by cells, which regulates their own functions as well as functions of their neighboring cells, causes spatial variability in the concentration of solutes in the culture media. This concentration gradient initiates convective mass transport and causes rapid distribution of the molecules over the entire volume of the media. As a result, autocrine and paracrine signaling is, at least temporarily, impaired [7] (Note: autocrine and paracrine signalings refer to the binding of a signal molecule to receptors on the surface of the same cell that secrets it and the neighboring cells, respectively, to initiate a cascade of intracellular and intercellular events). It is not difficult to think of other examples that highlight differences between the environmental conditions of a cell in vivo and in vitro. From the viewpoint of the pharmaceutical industry, the microenvironment disparity has significantly hindered the development of new drugs based on tests conducted with conventional cell-based assays to screen drug safety and efficacy.
188
Tavana, Chung, Kuo, and Takayama
Macroscale analytical techniques for biochemical analyses of cells also have several major drawbacks: These techniques mainly rely on large and expensive bench-top equipment, which makes them incompatible with point-of-need applications. They often require a large number of cells, which is not always available (such as in the study of precious cells and cell samples from human or animal models). In the case of single cell analysis, existing macroscale analytical techniques are slow and only a small number of cells can be analyzed per day [8]. The manually intensive nature of these methods is a major reason for their low throughput. Enhanced fabrication capabilities of the microelectronics industry in the early 1980s led to the advent of microelectromechanical systems (MEMS) that integrate mechanical components with electronics to create functional devices [9]. The technology was extended later to the manipulation of fluids inside micrometer scale channels and wells [10,11] and thus, the field of microfluidics emerged. A major source of motivation for microfluidics came from developments in different areas of biology such as molecular biology that required analytical methods with high throughputs and resolution not achievable previously with conventional methods. The commensurability of the linear dimensions of eukaryotic cells (10-100 µm) with microfluidic settings opened up the opportunity to use this technology for cellular applications. Microfluidics not only enables carrying out the three main processes of separation, culture, and analysis of cells, it also enjoys many inherent advantages over the macroscale cell handling counterparts. The benefits of microfluidic technology for cellular applications are discussed below from physiological, biological, and economical perspectives.
6.1.1 Physiological advantages A glance at the human anatomy reveals a large number of networks that are microfluidic in nature. These networks exist at different hierarchical levels, from system level such as the nervous, the lymphatic, and the cardiovascular systems to the functional units of various organs such as nephrons in the kidney, the network of alveolar ducts and sacs and the corresponding blood capillaries in the lung, etc. (Fig. 6.1). It therefore makes intuitive sense to mimic such in vivo structures with appropriate microfluidic devices for in vitro cell-based studies. Microfluidic technology also makes it possible to recreate tissues of different organs on a single chip by culturing various types of cells at different sections of an interconnected network of channels and chambers [12]. Such “body-on-a-chip” (BOC)
Microfluidic Systems for Cellular Applications
189
platforms would allow communication between distant cells through endocrine signaling, revealing information that could not be learned from conventional cultures. Endocrine signaling plays an important role in regulating the response of the human body to various pharmaceuticals and chemicals. Therefore, functional BOCs can provide realistic and physiologic platforms for drug testing, reduce the need for animal-based tests and lower the cost of developing new pharmaceuticals. A more distant prospect of BOCs is customizing the treatment of patients with severe disorders. Cell samples from individual patients can be cultured in a chip and responses to particular drugs can be assessed to determine the most effective treatment. (a)
(b)
Terminal Bronchiole Branch of Pulmonary Vein Branch of Pulmonary Artery
Smooth Muscle
Respiratory Bronchiole
Alveoli Capillary
300 µm
Fig. 6.1 Many systems and functional units of the body are microfluidic in nature such as (a) the vessels of the lymphatic and cardiovascular systems in neck and head (Image reproduced, with permission, from 3D4Medical.com/Getty Images [13]) and (b) the respiratory bronchioles with clusters of alveoli and blood vessels (Image reproduced, from ref. [14] with permission from The McGraw-Hill Companies (©2006)
6.1.2 Biological advantages Microfluidic cellular systems are valuable tools from a biological standpoint too. First, both static and dynamic culture conditions can be realized in microfluidic devices. Given that convection is negligible in microchannels that are not perfused and diffusion is the operative mode of molecular transport, static cell cultures preserve autocrine and paracrine factors in the
190
Tavana, Chung, Kuo, and Takayama
proximity of cells. The formation of gradients of soluble factors in the channels is physiologically significant as it is a key regulator of the fate and the phenotype of various cells. Dynamic cell culture is an even more appealing feature of microfluidic systems since it enables the recreation of various physiologically-relevant processes. For example by controlling the flow rate in microchannels, cell-cell interactions through autocrine, paracrine signaling, and even endocrine effects can be systematically manipulated. With dynamic cultures, it is also possible to mimic a range of in vivo processes directed by mechanical stimulation of cells, such as the modification of morphology of endothelial cells under shear stress [15], rolling of leukocytes along the luminal surface of blood vessels under blood flow [16], and injury of small airway cells due to fluid mechanical stresses generated by propagation of liquid plugs during reopening of closed pulmonary airways [17]. Second, materials with tunable properties might be used for the fabrication of microfluidic devices and different geometrical features can be incorporated inside microchannels to investigate various aspects of cell-ECM interactions. For example, the effect of ECM elasticity on the fate of embryonic stem cells was studied by tailoring surface properties of ECM [18]. Some aspects of cell-ECM adhesion characteristics were revealed through measurements of forces exerted by cells on an array of microposts [19], and 3D scaffolds were incorporated in microchannels to provide cells with a more in vivo-like ECM environment [20,21]. Third, fluid flow in microchannels is almost always laminar with typical Reynolds numbers of less than one. There is no appreciable mixing between streams of fluids other than by diffusion. Laminar flow is one of the key features of microfluidic systems and is exploited for a wide range of on-chip applications including subcellular positioning of small molecules [22], protein fractionation [23], and rapid immunoassays [24]. Overall, laminar flow provides a gentle environment for separation, culture, and analysis of cells. Fourth, there is a growing emphasis in molecular biology on single cell analysis to understand phenotypic heterogeneity among cells in a population. Microfluidic systems are advantageous since they can enhance the throughput of single cell analysis and integrate various cell handling and processing steps [25].
Microfluidic Systems for Cellular Applications
191
6.1.3 Economical advantages The application of microtechnology to cell biology can realize significant economical advantages over conventional cell handling processes. Microfluidic devices have small volumes and large surface area-to-volume ratios. As a result, only small volumes of reagents (several nanoliters) are consumed and fluidic reaction times are significantly shortened. Minute reagent volumes also reduce the amount of chemical waste produced. The process of device fabrication can be straightforward and commonly-used polymer materials for fabrication, e.g. poly(dimethylsiloxane) (PDMS), are relatively inexpensive (e.g. compared to silicon microdevices). Furthermore, high throughput microfluidic platforms can be realized by including many copies of a design on a single chip to simultaneously test the effect of different samples, say of a drug, on cells. It is also possible to integrate all three steps of cell separation, culture, and analysis into a single chip. Integrated devices can be complex and challenging to operate but when successful, reduce processing cost and time. The small footprints and low power requirements also make them suitable for portable, point-of-need applications. The discussion above outlined the shortcomings of conventional techniques for cell studies and highlighted microfluidic technology as a useful alternative. In the next section, although far from exhaustive, a selection of novel microfluidic platforms developed for separation, culture, and analysis of cells is presented. Then, commercialized microfluidic devices for cellular applications will be briefly reviewed. Finally, the chapter will conclude with future prospects of microfluidic cell systems.
6.2 Microfluidic Technology for Cellular Applications 6.2.1 Microfluidic cell isolation/separation Separation and isolation of cells is a key process for a variety of applications including: Fractionating whole blood samples to isolate white and red blood cells, deriving helper T lymphocytes to monitor HIV treatments, capture of rare circulating tumor cells to identify key biological determinants of blood-borne metastases [26], and deriving a stem cell line from mature tissues or liposuction aspirates. Commercially available cell separators often have several limitations such as large foot-print, the need for expensive lasers and detectors as well as trained operators.
192
Tavana, Chung, Kuo, and Takayama
Microfluidic systems are low-cost and easy-to-use alternatives. Downscaling cell separation systems to dimensions comparable to that of cells and harnessing unique fluidic properties at the micrometer scale results in increased speed, reduced sample consumption, smaller foot-print, and higher throughput. Separation mechanisms based on dielectrophoresis, capillary electrophoresis, acoustic forces, and optical traps can all be incorporated into microfluidic devices [27-34] and optical components can be integrated into these platforms. Below, a number of microfluidic cell separation methods are described. The selection highlights the flexibility of microfluidic systems to employ physical and biological principles for separation/isolation of different cell types including precious cells. Simplicity of fabrication, small reagent volumes, gentle separation, fast separation time, and user-friendliness are common features of these settings. 6.2.1.1 Microfabricated fluorescence-activated cell sorter (µFACS)
Conventional fluorescence-activated cell sorters (FACSs) are widely used in clinical medicine and biological sciences. FACSs are remarkably efficient, however they are costly, mechanically complex, difficult to sterilize, and require large sample volumes as well as trained personnel for operation/maintenance. Inexpensive devices that rapidly sort cells and particles would facilitate screening of cell populations and combinational chemistry libraries. Fu et al. developed an elastomeric microfabricated FACS (µFACS) based on electroosmotic flow [35]. The device resembled a T-junction with inlet, collection, and waste chambers. Two sorting algorithms were used: “forward” and “reverse”. In the “forward” method, cells flowed in a buffer solution from the input channel to the waste channel. If the fluorescent intensity of a cell was above a preset threshold, the voltages were temporarily changed to divert the cell to the collection channel. The “reverse” method was used to identify rare cells. Cells were moved at a high speed from the input to the waste. The flow was stopped when a cell of interest was detected. By the stoppage time of the flow, the cell was past the junction toward the waste channel. The system was then run backward at a slow speed from the waste to the input, and the cell was switched to the collection channel when it passed through the detection region. The device was used to sort wild E. coli cells from GFP-expressing E. coli cells. An enrichment of 80- to 96-fold and a sample throughput of 10-20 cells/sec was
Microfluidic Systems for Cellular Applications
193
achieved. Later, an optimized device from this group gave a two fold increase in the throughput [36]. However, this was still several orders of magnitude lower than that of conventional FACS, i.e. 10000-20000 cells/sec. To reduce the efficiency gap between conventional and microfabricated FACSs, Wolff et al. designed a pressure-driven µFACS silicon chip with three inlets and two outlets [37]. A sample of chicken red blood cells (CRBCs) and fluorescent latex beads was introduced from a middle inlet and sheathed by buffer streams from two sides. The fluorescent beads were excited near the junction of collecting and waste channels. The fluorescence signal was detected by a photomultiplier tube and the system activated a valve to force cells of interest to the collecting channel. A 100-fold enrichment of fluorescent beads was achieved at a throughput of 12000 cells/sec, i.e. 1000 times higher than that of previous µFACS. In a separate design, a different configuration was used to generate hydrodynamic focused sample flow. The sample was introduced into the buffer stream through a “chimney” like structure, being carried downstream the channel like a thin wisp of smoke from a chimney. This ensured lower variations in the fluorescent signal, which is a common problem in devices with hydrodynamic focusing of sample using side sheath flows. Other interesting features of this chip were the integration of a chamber for holding and culturing the sorted cells and optics for the detection of cells on chip. Such devices are promising signs for developing fully integrated functional microfluidic platforms for cell sorting, culturing, and analysis. Wang et al. described a µFACS that utilizes optical forces to sort mammalian cells [38]. A sample of Hela cells stably transfected with histone-GFP mixed with non-expressing parental Hela cells flowed from inlet and were sheathed to a narrow stream by hydrodynamic focusing using side flows. Cells in the focused flow first passed through an analysis section and then through an optical switching region of the device. Flow to the two output channels was asymmetrically biased and with no optical beam present, all cells moved to the waste channel. When a target cell was detected, the optical switch was activated and a focused laser spot deflected the cell to the target output channel. Because this switch operates by displacing the cell within the laminar flow rather than by making a transient change to the fluid flow, switch rates can be high. Throughputs of up to 106 cells/sec, recovery rates of >85%, and collection well purities of up to ~83-99% were obtained. Cell viability was evaluated by examining the expression levels of two cellular stress indicator genes HSPA6 and FOS. The results
194
Tavana, Chung, Kuo, and Takayama
showed no increase in the expression levels of these genes, an indication of viable cell populations. µFACSs have been optimized to yield reasonably high enrichments and throughputs. These settings are particularly useful for sorting of precious cells and applications where only a small number of cells is available, e.g. 100-100,000 cells, which cannot be handled with conventional flow cytometers. In addition, µFACSs offer many other advantages including consumption of small volumes of reagents, low cost, easy device sterilization, minimum sample carryover, and potential to integrate various other functional units to realize a fully self-contained lab-on-a-chip platform. 6.2.1.2 Microfluidic magnetic cell separation
Cells can be isolated under the influence of a magnetic field. Magnetic cell separation (MCS) is divided into two classes [39]: (i) separation using native magnetic susceptibility of cells and (ii) separation using antibodycoated magnetic beads. The former technique relies on the property of certain cells such as deoxygenated red blood cells to contain paramagnetic material [40-42]. Paramagnetic cells can be separated from the cell suspension under a magnetic field. Microfluidic devices that use this property to separate a cell type of interest contain a small ferromagnetic wire loosely packed into a region near the poles of a large magnet. The small wires generate high magnetic field gradients when magnetized and trap the paramagnetic cells. Han et al. demonstrated continuous separation of red blood cells in a microfluidic device by passing whole blood around a 50 µm magnetized nickel stripe [43]. Cell separation can also be realized through specific binding of antibody-coated magnetic beads with antigens expressed on a cell surface [44,45]. A number of microfluidic platforms have been described for the separation of magnetic beads [46-48] and cells [49-51] using this technique . For example, Inglis et al. incorporated magnetic strips into a silicon substrate and magnetized them by applying an external field [49]. The strips were placed at an angle to the direction of fluid flow carrying narrow streams of cells over the stripes. Magnetically labeled cells were attracted to the strips, whereas unlabeled cells did not interact with the strips and followed the direction of fluid flow. Separation of leukocytes from whole human blood was demonstrated. MCS is an inexpensive method for the separation and enrichment of different cell types including tumor cells. However, preventing permanent adhesion of the magnetically labeled cells to the strips is a great challenge.
Microfluidic Systems for Cellular Applications
195
Normally, a significant number of magnetically labeled cells (~50%) either permanently stick to the nickel strips or are not sufficiently attracted to the strips to be separated [39]. 6.2.1.3 Transient cell-ligand adhesion
Cells may be isolated by recruiting natural, physiological process of leukocyte adhesion to blood vessel walls at the sites of injury or inflammation. For example during an inflammatory response, glycoproteins of white blood cells (WBCs) are recognized by cell adhesion molecules, including the selectin family, on the surface of endothelial cells (ECs) lining blood vessels at the inflammation site. This recognition causes WBCs to adhere to the blood vessels. This transient adhesion between WBCs and surface ligands retards movement of the cells. Due to a continuous dissociation/formation of bonds between proteins on the surface of WBCs and ECs under shear from the blood flow, WBCs roll slowly along the blood vessels luminal surface and eventually, migrate to the inflamed tissue. Chang et al. used this principle to capture and concentrate cells in microfluidic devices [16]. Channels contained arrays of either square or offset micropillars with spacing larger than the size of the cells to allow for free motion of the cells. Channels were treated with human E-selectin IgG as the adhesion protein. A suspension of HL-60 cells bearing ligands for Eselectin was flowed at 1 µl/min. In the first minute of the experiment, channels with offset pillars captured about 95% of the cells over a distance of 700 µm whereas the capture efficiency of the square array design was about 72%. With increasing the flow time, the efficiency of the square geometry improved to 85-94%. Both designs were capable of enriching cells several hundred folds from the original concentration. This technique is potentially applicable to chromatographic fractionation of adherent cells because different cell types have different affinities for a given adhesion molecule. Separation of HL-60 and U-937 myeloid cells was demonstrated, albeit with low resolution. A similar strategy was applied by Nagrath et al. to separate circulating tumor cells (CTCs) from peripheral whole blood samples in a single step without a need for pre-dilution, pre-labeling or other sample processing steps [26]. CTCs are rare and comprise as few as one cell per 109 haematologic cells in the blood of patients with metastatic cancer. The separation is mediated by the interaction of target CTCs with anti-epithelial cell adhesion molecule (EpCAM)-coated array of microposts incorporated in the channels. Overexpression of EpCAM in the carcinomas of lung, breast,
196
Tavana, Chung, Kuo, and Takayama
and prostate provides the specificity for CTCs capture from the nonfractionated blood. CTCs were identified in 99% of the blood samples donated by cancer patients. The purity of CTCs capture varied between ~50% to ~70% for cancers of different origin and a capture specificity of 100% was achieved. 6.2.1.4 Laminar flow
Approximately 10% of couples have infertility problems [52]. The most advanced treatment for male-related infertility is an in vitro fertilization (IVF) technique known as intracytoplasmic sperm injection (ICSI) [53], where an oocyte is fertilized by the direct injection of a sperm. To maximize the rate of successful pregnancy and birth, it is important to select the most viable sperm. Although ICSI has significantly reduced the number of viable sperm required for fertilization, it bypasses all sperm selection processes. Due to the limitations of current techniques such as centrifugation and swim-up processes [54] for the isolation and selection of viable sperms, doctors frequently resort to hand sorting through dead sperm and debris, a procedure that can take hours in some cases. Developing novel processes to isolate the most viable sperm would greatly benefit IVF processes in clinical settings. Cho et al. described a PDMS microscale integrated sperm sorter (MISS) that separates motile sperm from non-motile sperm and cellular debris [55] (see Fig. 6.2). Horizontally oriented reservoirs with different heights at inlets and outlets provided a passively driven gravity pumping system and facilitated liquid movement in the channels. Separation in this microfluidic system relied on the existence of multiple parallel laminar streams with no mixing at the interface between them. Non-motile human sperm, ~60 µm in length, and particles diffused slowly and remained within their initial streamlines. In contrast, motile sperm were very mobile and swam at ~20 µm/sec and distributed randomly within the 500 µm-width channel. A bifurcation at the end of the separation channel was used to collect motile sperm that deviated from its inlet streamlines (Fig. 6.2). The purity of motile sperm after sorting was nearly 100%, regardless of motile sperm purity prior to the sorting. The yield of the MISS, i.e. the ratio of the number of collected motile sperm to the total number of motile sperm at the inlet reservoir was ~40%, which is comparable to the recovery rates of conventional sperm sorting methods [54,56]. In principle, this separation mechanism is similar to that of “H-filter” where rapidly diffusing small molecules exit through a different outlet from large molecules and particles that diffuse slowly [57,58]. The main difference is that the MISS takes
Microfluidic Systems for Cellular Applications
197
advantage of active movement of cells whereas the H-filter relies on the passive diffusion of molecules.
Fig. 6.2 Three-dimensional illustration of the microfluidic channel design with horizontally oriented fluid reservoirs and separation of motile sperm from nonmotile sperm (Reprinted with permission from [55] Copyright (2003) American Chemical Society.).
The MISS has several advantages over existing techniques: (i) it is a selfcontained functional microdevice that manipulates cells without a need for electronics and external power sources, (ii) it offers an efficient way for the isolation of sperm from samples with few motile sperm that are difficult or impossible to process using conventional methods, (iii) it facilitates a mild biomimetic process and avoids centrifugation and sample compaction that are known to cause sublethal damage to sperm [54,59], (iv) it is small, portable, easy to use, disposable, and inexpensive. The MISS is promising for use in large scales in clinical settings and opens up the possibility of making self-contained bioassays for convenient at-home screening of male infertility.
198
Tavana, Chung, Kuo, and Takayama
6.2.1.5 Deterministic hydrodynamic
A microfluidic device was described by Davis et al. that separates WBCs and red blood cells (RBCs) from whole blood and fractionates different WBCs (lymphocyte, monocyte, granulocyte, T cell, B cell) according to their size [60]. WBCs are more or less spherical and range from 5 to 20 µm in diameter, whereas RBCs are disk-shaped with 8 µm in diameter and have a 2 µm thickness. The principle of operation is as follows: A cell suspension flows through an array of microposts incorporated in the channel. Each row of posts is slightly offset laterally with respect to the row immediately above it. Particles smaller than a critical hydrodynamic diameter (Dc) follow streamlines cyclically through the gaps, moving in an average downward flow direction. Particles larger than Dc do not fit into the first streamline and bump into a neighboring streamline at each post. Thus they move at an angle with respect to the direction of fluid flow, determined by the ratio of post offset to the row-to-row spacing. Fig. 6.3 illustrates the design. Whole blood and a buffer solution entered into different channels at the top of the device and maintained a vertical laminar flow throughout the device. The active region consists of 13 consecutive regions with various Dc of 3 to 9 µm. Cells smaller than 3 µm moved straight down, those larger than 9 µm bumped at all regions of the device, and cells of intermediate size started by bumping in the upper regions of the device and eventually moved straight after entering a region with a Dc larger than the cell size. The bumped cells were diluted into the buffer solution. The device had three separate output channels to collect cells of slightly differing hydrodynamic radii. Cells behaving as objects larger than the maximum Dc for all 13 regions (>9 µm diameter) were thus displaced from the left edge to the right edge, exiting from channel 3, whereas those with a diameter of less than 3 µm passed straight down the device and exited from channel 1. The analysis of collected cells showed that more than 99% of all WBCs were displaced into channels 2 and 3 and the distribution was dependent on the cell size. For example, the ratio of lymphocytes in channel 2 to channel 3 was ~100 to 1. The total number of RBCs at the three outlets was also determined as 3×106 with 99% of the RBCs exiting from channel 1.
Microfluidic Systems for Cellular Applications
199
Fig. 6.3 Microfluidic device to separate WBCs from RBCs and platelets. Arrows indicate the vertical and bumping directions (Image reproduced, with permission, from ref. [60] Copyright (2006) National Academy of Sciences, U.S.A.).
6.2.1.6 Free flow acoustophoresis
Acoustic forces are widely used to separate suspended particles from their medium or from other particles [61,62] as well as to trap particles [63,64]. Acoustic separation and trapping systems are gentle to biological systems and also easy to operate. Petersson et al. described a microfluidic platform that utilizes the combination of laminar flow and acoustic forces to separate particles and cells based on their size and density in an ultrasonic wave field [32]. The technique is called free flow acoustophoresis (FFA). Suspended particles/cells entered a 350 µm-wide channel through two side inlets. The medium without suspended particles flowed through the center inlet into the main channel. An acoustic force field perpendicular to the flow direction was generated between the side walls of the channel using a piezo ceramic actuator. As the particles/cells moved along the flow path, the acoustic force displaced them toward the center of the channel. By balancing the flow rate, acoustic force, and particle mixture, a particle gradient developed across the channel. Due to the laminar flow in the channel, this gradient was further fractionated down the channel into several outlets. Fractionation of suspension of particles/cells with fairly similar densities was more challenging and required the addition of small amounts of cesium chloride to the medium to manipulate its density. The strategy was proved useful to fractionate a suspension of RBCs and platelets containing 1% by volume from each cell type in a device with two active outlets. 92% of the RBCs and 99% of the platelets were directed to the first and the sec-
200
Tavana, Chung, Kuo, and Takayama
ond outlets, respectively. A more complex cell suspension, i.e. Buffy coat, was also fractionated by FFA. Blood component therapy for transfusion, i.e. the use of pure RBCs, platelets, and leukocytes, as well as in modern transfusion therapy based on leukocyte-depleted platelets may benefit from such microdevices.
6.2.2 Microfluidic cell culture In vitro culture of animal cells has a wide range of applications. For example: to study the normal physiology or biochemistry of cells such as cell metabolism, to investigate the sequential or parallel combination of different cell types to generate artificial tissues, to evaluate the safety and efficacy of new pharmaceuticals, to synthesize valuable biologicals such as therapeutic proteins from large scale cell cultures, etc. Standard culture systems that are often performed on 2D treated plastic surfaces do not mimic the in vivo microenvironment of cells: oxygen tension is too high, the concentration of soluble factors is abnormally high, cell-cell interactions are unorganized, and 3D cues specific of the native tissue are absent [25]. Microfluidic technology provides the opportunity to create a physiologic cellular microenvironment by regulating the transport of nutrients and other biochemicals in microchannels in a spatio-temporal manner and engineering the ECM using different materials and geometries to mimic complex ECM in vivo and to provide cells with appropriate extracellular cues. The application of microfluidics to the culture of mammalian cells is discussed below in the context of a few examples. These studies demonstrate the feasibility of creating physiologic tissue models through controlled manipulation of the microenvironment of cells, an achievement that could not be realized previously with conventional dishbased cultures. 6.2.2.1 Recreation of pulmonary airways on a chip
The branching network of pulmonary airways is lined with a viscous liquid film secreted by airway epithelial cells. The primary function of airways is directing airflow to the sites of gas exchange with the blood. Many respiratory diseases such as chronic bronchitis, emphysema, and asthma are accompanied by overproduction of mucus in airways and often, by dysfunction of pulmonary surfactant. As a result, the free surface of the thick mucus-containing liquid film grows over time and eventually forms a liquid plug across the airway lumen. The liquid plug obstructs airflow, result-
Microfluidic Systems for Cellular Applications
201
ing in poor ventilation of the lung and insufficient blood oxygenation. Upon inhalation, the liquid plug propagates along the airway and eventually ruptures, reopening the occluded airways (Fig. 6.4a). This process is known as “airway closure and reopening” [65,66]. Movement of liquid plugs and reopening of occluded airways is known to generate large fluid mechanical stresses [67], which are suspected to damage small airway epithelial cells (SAECs). Nevertheless, the complexity of pulmonary fluid flows and the small size of airways (several hundred micronmeters) have hindered experimental study of the cellular level injury in vitro. Huh et al. developed a three-dimensional microfluidic device that closely mimics the microenvironment of airway epithelial cells in vivo [17]. The microfluidic airway system consists of two PDMS compartments separated by a polyester porous membrane (Fig. 6.4b). The upper and lower compartments represent the apical and basal chambers of airway epithelium, respectively. Primary human SAECs were seeded into the upper chamber and cultured on the membrane, which mimics the basement membrane. To promote the growth of the cells, both chambers were perfused with culture media for about one week. Once a confluent monolayer of cells was formed, the apical surface of the cells was exposed to an air-liquid interface to induce cellular differentiation. Maintaining the microfluidic culture system under this condition for about two weeks induced cellular differentiation, causing SAECs to express morphological and secretory phenotypes found in native airway tissues. Differentiation of SAECs was confirmed by immunoassay of Clara cell secretory protein as well as by immunohistochemical detection of tight junctions between epithelial cells. After three weeks of microfluidic culture, over 85% of cells were still alive. The microfluidic airway system was then integrated with a computerized liquid plug generator to study cellular damage during airway reopening. Preliminary results showed that plug propagation and rupture is deleterious to SAECs and cellular viability decreases significantly with increasing number of propagation and rupture events. The injurious effect of airway opening on lung epithelial cells was also reported by Bilek et al. using the propagation of a semi-infinite bubble in a narrow fluid-occluded channel lined with cells [68]. Such studies can enhance current understanding of cellular responses to complex mechanical forces of the pulmonary system and contribute to the design of strategies for preventing/treating lung injuries.
202
Tavana, Chung, Kuo, and Takayama
(a)
(b)
Air flow
Rupture
Fig. 6.4 Schematic of (a) occluded pulmonary airways and the reopening process, and (b) micro-engineered compartmentalized biomimetic small airway system (Image reproduced, with permission, from ref. [17] Copyright (2007) National Academy of Sciences, U.S.A.)
6.2.2.2 Liver on a chip
Many drug candidates fail clinical trials either because they damage the liver directly or because liver metabolites are toxic [69]. The problem stems, at least in part, from the inadequacy of standard cell-based assays for screening drug safety and efficacy. Development of new drugs and the process of drug discovery would greatly benefit from in vitro cell culture systems that can maintain the physiological liver functions of hepatocytes. However, the development of reliable liver cell cultures as “biosensors” for drug toxicity is quite challenging mainly due to the difficulty in maintaining differentiated phenotypes over prolonged periods. Dordick and co-workers developed a 3D cell culture array for highthroughput toxicity screening of drug candidates and their metabolites against different cell types [70,71]. A microarray of MCF7 breast cancer cells encapsulated in a hydrogel matrix (collagen or alginate) was spotted onto a functionalized microscope glass slide and then was incubated in culture medium for six days. To evaluate the response of cells to different cytotoxic compounds, this slide was stamped against a microarray of various toxins immobilized on a second glass slide. Cells were grown for three days prior to performing the viability assay. Griffith and co-workers developed a microbioreactor for the in vitro culture of liver tissue under continuous perfusion [72]. The bioreactor is integrated with a 3D scaffold that plays the ECM role. The scaffold consists of
Microfluidic Systems for Cellular Applications
203
a thin silicon sheet permeated with a regular array of microchannels from top to bottom. This scaffold sits on top of a microporous filter, which in turn is mechanically supported by a second scaffold. The reactor dimensions combined with the perfusate flow rates meet estimated values of cellular oxygen demands and provide shear stresses below physiological range. The channel walls were covered with a fluoropolymer during deep reactive ion etching (DRIE) of the silicon wafers and were subsequently coated with collagen I to enhance the adhesion and guide the morphogenesis of hepatocyte cells. Hepatocytes were seeded both in form of single cell suspension and pre-aggregated multicellular spheroids. In both cases, tissue-like structures formed inside the reactors. However, the microtissues formed from single cell suspension remained stable for four days only and then a progressive loss of structure was observed. On the other hand with hepatocyte spheroid seeding, the architecture and viability of tissues were preserved for up to two weeks and cultures exhibited cell-cell contact such as tight junctions and desmosomes, found in tissues in vivo [69,72]. Hepatotoxicity studies will benefit from the availability of such microengineered systems to investigate in vivo physiology and pathology in an in vitro environment. It has been observed that co-culture of hepatocytes with other cell types, including liver epithelial cells and Kupffer cells, prolongs the survival of cultured hepatocytes for up to two months and helps maintain liverspecific properties such as albumin secretion [73]. Furthermore, it is known that nonparenchymal cells participate in pathophysiological responses including drug toxicity and infection in conjunction with hepatocytes [72]. These findings emphasize that more physiological liver cell cultures representing a wide range of liver functions in vitro might only be realized by including appropriate homo- and heterotypic cell-cell interactions and possibly by considering various aspects of the liver microarchitecture in vivo. For example, Bhatia and co-workers described a 24-well culture system with microscale architecture fabricated using soft lithography for human liver cells [74]. Each well contained 37 colonies of hepatocytes on 500 µm diameter collagenous patterns that were surrounded by 3T3 fibroblasts. Unlike pure cultures that rapidly lost their morphological features and liver specific functions, the micropatterened co-cultures were stable and maintained liver specific functions for several weeks, as confirmed by different tests including albumin secretion, urea synthesis, and gene expression levels of liver-specific genes.
204
Tavana, Chung, Kuo, and Takayama
6.2.2.3 Microfuidics for stem cell research
Stem cells (SCs) are capable of dividing extensively and developing into many different types of cells in the human body. SCs hold great promise for cell-based therapy of many diseases including heart, lung, and blood diseases, diabetes, Parkinson’s and Alzheimer’s diseases as well as the repair of injured tissues and organs [75]. The main challenge is to understand differentiation pathways of SCs into various cell types and control the behavior of SCs in culture. Cellular microenvironment is one of the key regulators of functions of SCs and their differentiation fate [76,77]. Unlike traditional dish-based culture techniques, microfluidic cultures of SCs offer the advantage of manipulating physical and chemical environment of SCs in a spatio-temporally controlled manner and setting up scalable and parallel experiments. Growth and differentiation of human neural stem cells (hNSCs) was explored in a gradient-generating microfluidic platform by Chung et al. [78]. Diffusive mixing in the laminar flow regime throughout a network of microchannels generated a gradient of a mixture of several growth factors (GFs). A cell culture area at the bottom of the device was exposed to this gradient. Continuous perfusion ensured autocrine and paracrine signaling was minimized. It was demonstrated that cell proliferation was proportional to the concentration of GFs, whereas differentiation to astrocytes was inversely proportional to GFs concentration. Preserving autocrine and paracrine factors is also possible if cell-cell communication via soluble factors is concerned [79]. Kim et al. described a microfluidic device for perfused culture of stem cells with logarithmic range of flow rates. Syringe-driven flow combined with a network of fluidic resistances generated logarithmic flow conditions through separate cell-culture chambers in a single device [80]. Murine embryonic stem cells (mESCs) were cultured and perfused with media at rates of 0.001 to 1.1 µl/min for four days. Morphology of mESCs was found to be strongly influenced by the flow rate. Cells at the slowest perfusion did not proliferate. By increasing the flow rate, mESCs formed larger but fewer numbers of colonies. Larger colony areas at higher flow rates were ascribed to increased nutrient delivery, increased waste removal, and increased removal of proliferation-inhibiting secreted factors. The device was further modified to accommodate both logarithmic perfusion rates and concentration gradients in one microfluidic chip, making it possible to simultaneously explore a wider range of biological activities.
Microfluidic Systems for Cellular Applications
205
Figallo et al. described a micro-bioreactor comprising arrays of wells for the culture of different cell types including hESCs [81]. Each well was independently perfused with culture medium through a network of microchannels. To accommodate the culture of hESCs in a 3D setting, the wells were coated with a thin layer of a photopolymerizing hydrogel. The device allowed for controlled growth of hESCs under different flow configurations and cell densities and facilitated their differentiation into vascular cells. Recently, Zhong et al. developed a microfluidic platform to extract total mRNA from a single hESC and synthesize cDNA on the same device with a mRNA-to-cDNA efficiency superior to that of the bench-top counterparts [82]. The device may be useful for understanding the normal and pathological development of human cells and tissues. The growing interest in the microfluidic-based stem cell research over the past decade has led to the development of platforms with high levels of control over soluble microenvironment of SCs. Integration of heterotypic cultures and 3D cues is expected to thrust this area of research further. 6.2.2.4 Braille display-based microfluidic cell culture platforms
Transport of fluids relies on various pumping and valving mechanisms such as syringe pumps [83], gravity-driven pumps [55], and hydrogel valves [84]. Unfortunately, it is difficult to integrate such components into microfluidic devices. Furthermore, due to the slow nature of mixing in the laminar flow regime, these mechanisms cannot provide mixing and delivery of different reagents at various combinations and concentrations that is required for certain in vitro cell studies. A microfluidic plaform that utilizes programmed movement of arrays of pins of a Braille display (BD) as integrated valves and pumps was developed by Gu et al. [85]. Channel replicas are sealed against a thin PDMS sheet of ~140 µm. The network of microchannels is aligned onto a grid of Braille pin actuators. Valving action is generated by movement of a Braille pin up and down. Upward movement of a pin deforms the PDMS sheet and constricts the channel section immediately above it, whereas retracting the pin opens the constricted channel. Synchronous movement of three pins (valves) results in a peristaltic pump that moves fluids in a pulsatile fashion. This computerized fluidic system enables rapid mixing of several streams, multiple laminar flows with minimal mixing between them, and segmented flow of immiscible fluids (Fig. 6.5).
206
Tavana, Chung, Kuo, and Takayama
Futai et al. improved the BD-based microfluidics by reformulating the culture media to eliminate the need for exogeneous CO2, incorporating a transparent heater in the device to locally heat the cell cultivation area, and stabilizing the device using a set of a monolithic aluminum plates and hold-down clamps [86]. Incubator-free culture of C2C12 myoblasts and MC3T3-E1 osteoblasts was demonstrated over a period of two weeks without medium exchange. The BD-based cell culture platform is suited for on-site cell culture applications due to its compactness, portablity, optically accessibility of the cell cultivation area, and control over fluidic transport.
Fig. 6.5 (Left) The configuration of Braille display-based microfluidic setup and (Right) Segmented flow of perfluorodecalin and red and green food dyes (Images reproduced, with permission, from ref. [85] Copyright (2004) National Academy of Sciences, U.S.A.)
6.2.2.5 Endothelium on a chip
Endothelial cells (ECs) line the inner surface of blood vessels and form an interface between blood flow in the lumen and the rest of the vessel wall. In vivo, ECs are subjected to hemodynamic shear stress of 5-20 dyn/cm2, which induces various cellular responses [87-89] including in the morphology of ECs such that cells become elongated and aligned in the direction of the blood flow [90,91]. Understanding the effect of mechanical forces such as shear stress on ECs is fundamental to the study of certain diseases such as thrombosis and atherosclerosis. A BD-based microfluidic endothelium culture platform was described by Song et al. to systematically study the influence of physiological levels of
Microfluidic Systems for Cellular Applications
207
shear stress on ECs morphology [15]. Microchannels with different contact areas were incorporated to allow for the displacement of different volumes of media at a given pumping frequency (shown as small pump and large pump in Fig. 6.6), generating various levels of shear stress in a single device. A unique feature of this system is the generation of pulsatile fluid flow that represents the nature of blood flow in the arterial vasculature. ECs are known to discriminate between pulsatile and non-pulsatile flows [91], and hence in vitro EC culture systems must generate pulsatile flows to find physiologic value. ECs response to shear stress was evaluated using the angle of orientation (AO) (the angle between cell’s major axis and the flow direction) and the shape index (SI) (dimensionless number that is 1 for a straight line and 0 for a circle). For shear stresses of <1 dyn/cm2, no change was detected in the ECs morphology over 24 hrs. On the other hand for ~9 dyn/cm2 of shear stress, AO decreased by 20° and SI showed a decrease of 0.21. Cells were also exposed to a range of shear stress from 312 dyn/cm2 using the large pump. This resulted in a systematic change of AO and SI with shear stress, indicating alignment and elongation of ECs in the flow direction.
Fig. 6.6 Schematic of fluidic structure and close-up of ECs grown within an individual compartment. Braille pins acting as valves create three individual cellular compartments. Solid and dashed arrows indicate the direction of flow generated by large and small pumps, respectively. (Image reproduced, with permission, from ref. [15] Copyright (2005) American Chemical Society)
208
Tavana, Chung, Kuo, and Takayama
6.2.3 Microfluidic cell analysis The ability to rapidly analyze cells and their constituents is important in cell biology from various respects, for example to understand enzymatically-governed intracellular processes. Such reactions, in which the concentration of analytes can change in a fraction of a second, take place in signal transduction pathways where kinases and phosphotases transduce an extracellular signal into a cellular response. To measure the concentration of analytes involved in these signaling cascades, cells must be rapidly lysed and enzymatic reactions should be stopped [8]. Traditional techniques for biochemical analysis of cells are generally slow and of lowthroughput, making them insufficient for such studies. On the contrary, microfluidic settings for cell analysis have reduced assay scale and offer many advantages over macro-scale analytical systems including consumption of very small quantities of samples and reagents, high resolution and sensitivity of separation and detection processes, short reaction time, controlled chemical microenvironment, low cost, and small footprints. It has been shown that using microfluidic devices, cells can be lysed in < 33 ms and contents of 7-10 cells can analysed per minute. This is 100-1000 times faster than a bench-top CE single cell analysis platform [8]. In general, every cell analysis process involves capturing/trapping, culturing, and screening of cells of interest. There are four different strategies most often used for on-chip cell analysis: Microarrays, tissue model, single cell analysis, and chemotaxis assays. These strategies are described below. It is noted that the literature of microfluidic cell analysis contains a substantial amount of work; however, the following discussion covers only a few representative examples. 6.2.3.1 Cell microarrays
Microarrays are capable of performing high throughput analysis to address several experimental conditions on a single chip with superior cell handling to well-based assays and less reagent consumption. Lee et al. designed a microbioreactor that resembles physiological tissue conditions. It contains a number of “C”-shaped rings to trap HeLa cells and an outer channel for “blood” flow to feed cells through diffusion [92]. An 8×8 array of chambers provided multiplexing nanoliter culture environments with uniform cell loading, shear stress and pressure elimination, and a stable fluidic control. The response of HeLa cells to varying concentrations of fetal bovine serum (FBS) was analyzed using this device. Wang et al. also developed a 24×24 microfluidic chamber array for cell cytotoxicity analy-
Microfluidic Systems for Cellular Applications
209
sis [93]. Each chamber contains eight semi-circle micro cell sieves for cell trapping. Three cell types (BALB/3T3, HeLa, and bovine endothelial cells) were sieved and seeded in this microfluidic system and then analyzed against a panel of five toxins (digitonin, saponin, CoCl2, NiCl2, acrolein). Using these scalable microfluidic platforms, cell seeding and toxin exposure can be carried out within a single device in a multiplexed format, enabling high-density parallel cytotoxicity screening while minimizing cell and reagent consumption. 6.2.3.2 Tissue model
There are three main challenges to constructing an in vitro tissue model with a physiological environment for cell analysis: (i) the control of cell density without causing damage to the cell membrane, especially cells with high volumetric density such as cancer cells, (ii) active mass transport by blood vessels, and (iii) passive transport by diffusion controlled by endothelial cells. It is extremely difficult to resolve these problems with traditional cell assays, whereas a network of microchannels and micropores can mimic endothelial barriers and blood vessels. Lee et al. developed two microfluidic tissue model systems for cell analysis [94,95]. The first device enables screening and high-throughput analysis of live cell cancer toxicity in a multiplexed tissue-like culture. Each microfabricated culture unit consisted of three functional components: a 50 µm wide cell culture pocket, an artificial endothelial barrier with pores of 2 µm, and a nutrient transport channel. A high density of cancer cells were maintained for over one week in a solid tumor-like morphology when fed with continuous flow. The toxicity profile of the anti-cancer drug paclitaxel was collected and analyzed on HeLa cells cultured for up to five days of continuous drug exposure. The second device mimicked an artificial liver sinusoid and was used for primary hepatocyte culture and screening the hepatotoxicity of a metabolism mediated liver toxicant, diclofenac. The hepatocyte culture chamber was separated from the nutrient transport chamber by a set of parallel microchannels with a cross section of 2 µm2 that formed an endothelial barrier. The endothelial barrier between culture area and the nutrient transport channel concentrated cells in the culture area and minimized convective flow through the cell culture region while allowing diffusive transport. Primary hepatocytes were maintained viable for 7 days in this platform. Diclofenac showed no adverse effect on primary hepatocytes in short term toxicity screening (4 h), but prolonged exposure (24 h) caused toxicity in metabolically active hepatocytes.
210
Tavana, Chung, Kuo, and Takayama
6.2.3.3 Single cell analysis
Single cell analysis creates new opportunities to investigate fundamental cell mechanisms such as cell signaling cascades, intracellular polarization, and activation of gene expression under external chemical stimuli. However, it is hard to analyze only one cell in conventional cultures such as Petri dishes and well-based assays. In microfluidic devices, an individual cell can be trapped and separated from bulk suspension by flow control or using tiny traps, and then the cell can be analyzed under the flow of specific stimuli in microchannels. Wheeler et al. developed a microfluidic system capable of gently isolating a single cell from the bulk cell suspension and delivering nanoliter volumes of reagents for cell reaction and analysis [96]. Multistep receptor-mediated calcium influx of Jurkat T-cells and U937 cells under the effect of ionomycin and some antibodies such as human IgG was detected and analyzed by fluorescence intensity using this device. Takayama et al. developed a microfluidic system, PARTCELL, with multiple laminar fluid streams to deliver different chemicals to defined regions of a cell (Fig. 6.7) [22]. Treatment of the right and left poles of a single capillary endothelial cell with Mitotracker Green FM and Mitotracker Red CM-H2XRos was demonstrated. Sawano et al. used laminar flows containing rhodamine-labeled epidermal growth factor (EGF) to locally stimulate single live COS cells and analyze parameters that permit ligand-independent lateral propagation of EGF signaling [97].
Fig. 6.7 Differential manipulation of regions of a single bovine capillary endothelial cell using multiple laminar flows in a microfluidic device [22].
Microfluidic Systems for Cellular Applications
211
6.2.3.4 Chemotaxis assays
Chemotaxis refers to migration of motile cells towards higher concentrations of chemoattractants. The most common conventional assays are Boyden chambers, Zigmond chambers [98], and Dunn chambers [99]. Although these methods are easy to use, they provide limited control over the gradient due to diffusion and yield only endpoint results with little information about complex cellular mechanism and movement during the chemotaxis process. Jeon et al. designed a microfluidic device to generate gradients of chemoattractants with high stability and control. The microfluidic system consists of an embedded network of microchannels with two main regions: the gradient generation region that is a pyramidal branched array of microchannels to split, mix and combine fluid streams as they flow through and the observation region in which cells are placed and analyzed. Both linear and nonlinear concentration gradients were realized. Chemotaxis of neutrophils [100,101] and breast cancer cells [102,103] has been demonstrated. Nonlinear gradients resulted in a better chemotaxis response than linear gradients.
6.3 Commercialization of Microfluidic Technology Microfluidics is still an adolescent technology; nevertheless, its initial transition from proof-of-concept demonstrations to successful commercialization is promising. Many microfluidic startup companies have been founded in the past several years and the technology has been presented now for a variety of applications, mainly in the realms of cell culture and analysis. The following review briefly discusses some of these applications. Each year in the pharmaceutical industry, billions of dollars in revenue are invested to research, develop and market new drugs. Studies show that discovering, developing and launching a new pharmaceutical may take several years and cost up to $2.0 billion depending on the therapy or the developing firm (including the cost of prospective formulations that fail) [104-107]. A fairly high portion of the expense is due to the inadequacy of traditional cell-based assays, or even animal models, that are routinely used to predict drug toxicity, pharmacokinetics, and metabolism in the early phase of drug discovery. As a result, a potentially toxic drug may pass this stage, but fail later in clinical trials. Given the expensive nature of the drug discovery process, novel cell culture systems that facilitate efficient testing and reliable screening of drug candidates would be invaluable
212
Tavana, Chung, Kuo, and Takayama
assets to pharmaceutical companies to reduce drug development cost and time. Realistic conclusions as to whether a new chemical entity is indeed safe (or toxic) may be drawn from such assays only if they provide cells with a microenvironment where target cells can interact with other cells, ECM, and soluble factors as they would in their native environment. It was discussed throughout this chapter that microfluidic systems recreate a physiologic-like microenvironment for cells. This notion has been pursued by several startup companies such as Hµrel (http://www.hurelcorp.com) and CellASIC (http://www.cellasic.com) to develop microfluidic cellbased assays for drug-related studies. The former company has developed a microfluidic circuit, “animal-on-a-chip”, that consists of separate but fluidically interconnected compartments, each containing a culture of cells of a specific tissue or organ. A culture medium that serves as a “blood surrogate” circulates in a series of microchannels connecting these compartments and allows interactions between multiple tissue types and pharmacologic compounds of interest added to the medium. Preliminary tests with cancer chemotherapeutic drugs tegafur and 5-FU in a device containing hepatocytes cultured in the liver compartment and colon cancer cells in the target tissue compartment revealed a dose-dependent cytotoxicity of the drugs to the target cancer cells, indicating that tegafur and 5-FU are metabolized to an active drug in the liver compartment. Interestingly, the same drugs were ineffective in a control static assay. CellASIC has designed a microfluidic platform to study the efficacy of anticancer drugs. The device contains a perfusion barrier that separates a cell culture area and a nutrient delivery channel. Microfluidic culture of the immortalized cervical cancer cell line, HeLa cells, resulted in a high density of cells, as observed in tumors in vivo. Exposing cells to the anticancer drug paclitaxel, which was added to the nutrient flow, caused cells to exhibit multicellular resistance, with properties similar to clinical tumor resistance. In contrast, exposure of Hela cells grown in a standard monolayer culture to the same drug resulted in significant cell death after 24 h [95]. These two examples demonstrate typical opportunities that microfluidic cell-based assays offer in the area of pharmacology. Intensive research is currently underway in several academic laboratories and companies to develop platforms that facilitate reliable screening of drug safety and efficacy. Another area of activity of microfluidic companies concerns infertility treatment. Traditional in vitro fertilization (IVF) procedures have fairly low success rates because IVF of ovum is performed in non-physiologic culture conditions and embryos are maintained in such an environment
Microfluidic Systems for Cellular Applications
213
prior to the IVF process. This procedure does not always produce healthy embryos. To increase the chances of fertilization and eliminate the need for expensive repeated trials, fertility clinics resort to transferring multiple embryos that sometimes results in multiple births. Such shortcomings of traditional IVF have motivated companies such as Incept BioSystems (http://www.inceptbio.com) and Vitae LLC (http://www.vitaellc.com) to develop microfluidic settings that generate physiologic culture conditions for cells during the entire fertilization process and therefore, enhance the rate of embryo development and maximize IVF success rates. Other companies active in the area of IVF are Strex (http://www.strex.jp), which offers a commercial version of the microfluidic sperm sorting (MISS) platform and Pria Diagnostics (http://www.priadiagnostics.com), which features a portable diagnostics system for male fertility screening. The recognition that microfluidic cell culture systems can closely resemble a physiologic microenvironment for cells has prompted few companies to commercialize them in various configurations as likely replacements for conventional culture dishes. For example, Ibidi (http://www.ibidi.de) has developed disposable platforms that enable the investigation of cell culture under static and continuous perfusion conditions, immunofluorescence, and chemotaxis. Cellix (http://www.cellixltd.com) has commercialized microfluidic cell-based assays to investigate cell adhesion, migration, invasion, chemotaxis and shear stress models in a number of physiological and pathological conditions such as angiogenesis, autoimmune diseases, and cardiovascular diseases. Bionas (http://www.bionas.de) features microfluidic chips to screen cellular metabolic activity using integrated sensors for the measurement of PH, oxygen content, etc. There are currently several other microfluidic companies commercializing assays for cellular applications: Micronics (http://www.micronics.net) and Agilent Technologies (http://www.home.agilent.com) market microflow cytometers, Caliper Life Sciences (http://www.caliperls.com) feature cellbased assays to evaluate the impact of various compounds on the immune system and to test synergistic effects of drug combinations, Gyros (http://www.gyros.com) offers microfluidic immunoassays on a compact disc that enable the analysis of a large number of samples in a short time, and RainDance Technologies (http://www.raindancetechnologies.com) offers droplet-based microfluidic assays for applications such as FACS and RNAi screening. The increasing trend in the number of microfluidic companies as well as in the range of available products is expected to continue over the next years.
214
Tavana, Chung, Kuo, and Takayama
6.4 Concluding Remarks With only over a decade from its emergence, microfluidic technology has come a reasonable way. The technology has not yet lived up to the early expectations for widespread commercialization, but development is active and expanding. New capabilities have been introduced into the area of cell biology and tremendous advances have already been realized at the proofof-concept level. State-of-the-art microfluidic platforms harness fluidic properties at the microscale to facilitate culture, separation/isolation, and analysis of mammalian cells. While the latter two branches of on-chip cell studies have undergone a rather smooth progress and various techniques have been devised to isolate and analyze cells, in vitro culture of cells with the aim of mimicking physiological conditions is not as straightforward. The challenge is to recreate all critical aspects of the intricate microenvironment of cells in an in vitro culture system where cells interact with soluble signaling molecules, ECM, and other cells as they would do in their native tissue. More progress in this respect can expedite the transition of microfluidics from the demonstration stage to targeted applications. Involvement of biologists is critical for resolving this multifaceted problem. Active participation in cross-disciplinary meetings and close collaboration between engineering scientists and biologists can make this happen. Other important goals to achieve are to automate, integrate and parallelize complex tasks into complete, functional systems for greater accuracy and reproducibility, smaller sample sizes, and higher throughput. For example, an integrated microfluidic platform could facilitate on-chip separation of a cell type of interest, delivery of the cell to a lysis chamber and subsequently to multiplexed sandwich immunoassays using integrated optics or label-free detection [25]. Such highly integrated cell-based microfluidic devices will find great applications in biomedical and pharmaceutical research as well as in clinical settings and home-test systems as point-of-care devices. The impact of microfluidic systems will become more evident if they are utilized by a wide range of users, e.g. cell biologists, clinicians, and environmental engineers, for daily routine tasks. This can be realized if functional microfluidic systems meet key criteria such as user-friendliness and robustness and comply with pre-defined quality control standards to ensure reliability of operation and reproducibility of analysis.
Microfluidic Systems for Cellular Applications
215
References 1. (1998) Cell separation methods and applications. Recktenwald D, Radbruch A (eds) Marcel Dekker, Inc., New York. 2. Lindahl PE (1956) On the counterstreaming centrifugation in the separation of cells and cell fragments. Biochim. Biophys. Acta 211:411-415. 3. Boyum A (1974) Separation of blood leukocytes, granulocytes and lymphocytes. Tissue Antigens 4(4):269-274. 4. Wachtel SS, Sammons D, Manley M, Wachtel G, Twitty G, Utermohlen J, Phillips OP, Shulman LP, Taron DJ, Muller UR, Koeppen P, Ruffalo TM, Addis K, Porreco R, MurataCollins J, Parker NB, McGavran L (1996) Fetal cells in maternal blood: Recovery by charge flow separation. Human Genetics 98(2):162-166. 5. Smeland EB, Funderud S, Kvalheim G, Gaudernack G, Rasmussen AM, Rusten L, Wang MY, Tindle RW, Blomhoff HK, Egeland T (1992) Isolation and characterization of human hematopoietic progenitor cells-an effective method for positive selection of Cd34+ cells. Leukemia 6(8):845-852. 6. Harrison RG (1910) The outgrowth of the nerve fiber as a mode of protoplasmic movement. J Exp Zool 9:787-846. 7. Yu H, Meyvantsson I, Shkel IA, Beebe DJ (2005) Diffusion dependent cell behavior in microenvironments. Lab Chip 5:1089-1095. 8. McClain MA, Culbertson CT, Jacobson SC, Allbritton NL, Sims CE, Ramsey JM (2003) Microfluidic devices for the high throughput chemical analysis of cells. Anal Chem 75:5646-5655. 9. Petersen KE (1982) Silicon as a mechanical material. Proc IEEE 70:420-457. 10. Harrison DJ, Fluri K, Seiler K, Fan Z, Effenhauser CS, Manz A (1993) Micromachining a miniaturized capillary electrophoresis-based chemical analysis system on a chip. Science 261:895-897. 11. Jacobson SC, Hergenrsder R, Koutny LB, Ramsey JM (1994) High speed separations on a microchip. Anal Chem 66:1114-1118. 12. Viravaidya K, Sin A, Shuler ML (2004) Development of a microscale cell culture analog to probe naphthalene toxicity. Biotechnol Prog 20:316-323. 13. http://www.gettyimages.com. 14. Widmaier EP, Raff H, Strang KT (2006) Vander's human physiology: The mechanisms of body function. 10th edn. The McGraw-Hill Companies, Inc., 15. Song JW, Gu W, Futai N, Warner KA, Nor JE, Takayama S (2005) Computercontrolled microcirculatory support system for endothelial cell culture and shearing. Anal Chem 77:3993-3999. 16. Chang WC, Lee LP, Liepmann D (2005) Biomimetic technique for adhesionbased collection and separation of cells in a microfluidic channel. Lab Chip 5:64-73. 17. Huh D, Fujioka H, Tung Y-C, Futai N, Paine R, Grotberg JB, Takayama S (2007) Acoustically detectable cellular-level lung injury induced by fluid mechanical stresses in microfluidic airway systems. Proc Natl Acad Sci USA 104:18886-18891.
216
Tavana, Chung, Kuo, and Takayama
18. Engler AJ, Sen S, Sweeney HL, Discher DE (2006) Matrix elasticity directs stem cell lineage specification. Cell 126(4):677-689. 19. Tan JL, Tien J, Pirone DM, Gray DS, Bhadriraju K, Chen CS (2003) Cells lying on a bed of microneedles: an approach to isolate mechanical force. Proc Natl Acad Sci USA 100:1484-1489. 20. Liu H, Roy K (2005) Biomimetic three-dimensional cultures significantly increase hematopoietic differentiation efficacy of embryonic stem cells. Tissue Eng 11(1&2):319-330. 21. Lui H, Lin J, Roy K (2006) Effect of 3D scaffold and dynamic culture condition on the global gene expression profile of mouse embryonic stem cells. Biomaterials 27:5978–5989. 22. Takayama S, Ostuni E, LeDuc P, Naruse K, Ingber DE, Whitesides GM (2001) Laminar flows: Subcellular positioning of small molecules. Nature 411:1016. 23. Giddings JC, Yang FJ, Myers MN (1976) Flow-field-flow fractionation: A versatile new separation method. Science 24:1244-1245. 24. Hatch A, Kamholz AE, Hawkins KR, Munson MS, Schilling EA, Weigl BH, Yager P (2001) A rapid diffusion immunoassay in a T-sensor. Nat Biotechnol 19(5):461-465. 25. El-Ali J, Sorger PK, Jensen KF (2006) Cells on chips. Nature 442:403-411. 26. Nagrath S, Sequist LV, Maheswaran S, Bell DW, Irimia D, Ulkus L, Smith MR, Kwak EL, Digumarthy S, Muzikansky A, Ryan P, Balis UJ, Tompkins RG, Haber DA, Toner M (2007) Isolation of rare circulating tumour cells in cancer patients by microchip technology. Nature 450:1235-1239 27. Fiedler S, Shirley SG, Schnelle T, Fuhr G (1998) Dielectrophoretic sorting of particles and cells in a microsystem Anal Chem 70:1909-1915. 28. Li H, Zheng Y, Akin D, Bashir R (2005) Characterization and modeling of a microfluidic dielectrophoresis filter for biological species. J Microelectromechanical Systems 14:103-112. 29. Jacobson SC, R RH, Koutny LB, Warmack RJ, Ramsey JM (1994) Effects of injection schemes and column geometry on the performance of microchip electrophoresis devices. Anal Chem 66:1107-1113. 30. Fu L-M, Leong J-C, Lin C-F, Tai C-H, Tsai C-H (2007) High performance microfluidic capillary electrophoresis devices. Biomed Microdevices 9:405412. 31. Bazou D, Kuznetsova A, Oakley WT (2005) Physical environment of 2-D animal cell aggregates formed in a short pathlength ultrasound standing wave trap. Ultrasound Med Biol 31:423-430. 32. Petersson F, Aberg L, Sward-Nilsson A-M, Laurell T (2007) Free flow acoustophoresis: Microfluidic-based mode of particle and cell separation. Anal Chem 79:5117-5123. 33. Schmidt BS, Yang AH, Erickson D, Lipson M (2007) Optofluidic trapping and transport on solid core waveguides within a microfluidic device. Opt Express 15:14322-14334. 34. Wang MM, Tu E, Raymond DE, Yang JM, Zhang HC, Hagen N, Dees B, Mercer EM, Forster AH, Kariv I, Marchand PJ, Butler WF (2005) Microflu-
Microfluidic Systems for Cellular Applications
217
idic sorting of mammalian cells by optical force switching. Nat Biotechnol 23:83-87. 35. Fu AY, Spence C, Scherer A, Arnold FH, Quake SR (1999) A microfabricated fluorescence-activated cell sorter. Nat Biotechnol 17:1109-1111. 36. Fu AY, Chou H-P, Spence C, Arnold FH, Quake SR (2002) An integrated microfabricated cell sorter. Anal Chem 74:2451-2457. 37. Wolff A, Perch-Nielsen IR, Larsen UD, Friis P, Goranovic G, Poulsen CR, Kutter JP, Telleman P (2003) Integrating advanced functionality in a microfabricated high-throughput fluorescent-activated cell sorter. Lab Chip 3:2227. 38. Wang MM, Tu E, Raymond DE, Yang JM, Zhang H, Hagen N, Dees B, Mercer EM, Forster AH, Kariv I, Marchand PJ, Butler WF (2004) Microfluidic sorting of mammalian cells by optical force switching. Nat Biotechnol 23:8387. 39. Inglis DW, Riehn R, Sturm JC, Austin RH (2006) Microfluidic high gradient magnetic cell separation. J Appl Phys 99:08K101-1–08K101-3. 40. Melville D, Paul F, Roath S (1975) Direct magnetic separation of red cells from whole blood. Nature (London) 255:706-708. 41. Graham MD (1981) Efficiency comparison of two preparative mechanisms for magnetic separation of erythrocytes from whole blood. J Appl Phys 52:25782580. 42. Takayasu M, Kelland DR, Minervini JV (2000) IEEE Trans Appl Supercond 10:927-930. 43. Han K, Frazier B (2004) Continuous magnetophoretic separation of blood cells in microdevice format. J Appl Phys 96:5797-5802. 44. Nakamur M, Decker K, Chosy J, Comella K, Melnik K, Moore L, Lasky LC, Zborowski M, Chalmers JJ (2001) Separation of a breast cancer cell line from human blood using a quadrupole magnetic flow sorter. Biotechnol Prog 17(6):1145 -1155. 45. Owen CS, Sykes NL (1984) Magnetic labeling and cell sorting. Immunol Methods 73:41-48. 46. Pekas N, Granger M, Tondra M, Popple A, Porter MD (2005) Magnetic particle diverter in an integrated microfluidic format. J Magn Magn Mater 293:584-588. 47. Pamme N, Manz A (2004) On-chip free-flow magnetophoresis: Continuous flow separation of magnetic particles and agglomerates. Anal Chem 76:72507256. 48. Berger M, Castelino J, Huang R, Shah M, Austin RH (2001) Design of a microfabricated magnetic cell separator. Electrophoresis 22:3883-3892. 49. Inglis DW, Riehn R, Austin RH, Sturm JC (2004) Continuous microfluidic immunomagnetic cell separation. Appl Phys Lett 85:5093-5095. 50. Furdui VI, Harrison DJ (2004) Immunomagnetic T cell capture from blood for PCR analysis using microfluidic systems. Lab Chip 4:614-618. 51. Lee H, Purdon AM, Westervelt RM (2004) Manipulation of biological cells using a microelectromagnet matrix. Appl Phys Lett 85:1063-1065.
218
Tavana, Chung, Kuo, and Takayama
52. Mosher WD, Pratt WF (1991) Fecundity and infertility in the United States: Incidence and trends. Fertil Steril 56:192-193. 53. Palermo G, Joris H, Devroey P, Van Steirteghem AC (1992) Pregnancies after intracytoplasmic injection of single spermatozoon into an oocyte. Lancet 340:17-18. 54. Smith S, Hosid S, Scott L (1995) Use of postseparation sperm parameters to determine the method of choice for sperm preparation for assisted reproductive technology. Fertil Steril 63:591-597. 55. Cho BS, Schuster TG, Zhu X, Chang D, Smith GD, Takayma S (2003) A passively-driven integrated microfluidic system for separation of motile sperm. Anal Chem 75:1671-1675. 56. Florence LHN, Liu DY, Baker HWG (1992) Comparison of Percoll, miniPercoll and swim-up methods for sperm preparation from abnormal semen samples. Hum Reprod 7:261-266. 57. Brody JP, Yager P (1997) Diffusion-based extraction in a microfabricated device. Sens Actuators A 58:13-18. 58. Weigl BH, Yager P (1999) Microfluidic diffusion-based separation and detection. Science 283:346-347. 59. Aitken RJ, Clarkson JS (1998) Significance of reactive oxygen species and antioxidants in defining the efficacy of sperm preparation techniques. J Androl 9:367-376. 60. Davis JA, Inglis DW, Morton KJ, Lawrence DA, Huang LR, Chou SY, Sturm JC, Austin RH (2006) Deterministic hydrodynamics: Taking blood apart. Proc Natl Acad Sci USA 103:14779-14784. 61. Gröschl M (1998) Ultrasonic separation of suspended particles-Part I: Fundamentals. Acustica 84:432-447. 62. Hawkes JJ, Coakley WT (2001) Force field particle filter, combining ultrasound standing waves and laminar flow. Sens Actuators B: Chem 75:213-222. 63. Lilliehorn T, Simu U, Nilsson M, Almqvist M, Stepinski T, Laurell T, Nilsson J, Johansson S (2005) Trapping of microparticles in the near field of an ultrasonic transducer. Ultrasonics 43:293-303. 64. Bazou D, Kuznetsova A, Coakley WT (2005) Cell-cell contact and membrane spreading in an ultrasound trap. Ultrasound Med Biol 31:423-430. 65. Grotberg JB (2001) Respiratory fluid mechanics and transport processes. Annu Rev Biomed Eng 3:421-457. 66. Cassidy KJ, Halpern D, Ressler BG, Grotberg JB (1999) Surfactant effects in model airway closure experiments. J Appl Physiol 87:415-427. 67. Fujioka H, Grotberg JB (2005) The steady propagation of a surfactant-laden liquid plug in a two-dimensional channel. Phys Fluids 17:082102-082117. 68. Bilek AM, Dee KC, III DPG (2003) Mechanisms of surface-tension-induced epithelial cell damage in a model of pulmonary airway reopening. J Appl Physiol 94:770-783. 69. Sivaraman A, Leach JK, Townsend S, Iida T, Hogan BJ, Stolz DB, Fry R, Samson LD, Tannenbaum SR, Griffith LG (2005) A microscale in vitro physiological model of the liver: Predictive screens for drug metabolism and enzyme induction. Curr Drug Metab 6(6):569-591
Microfluidic Systems for Cellular Applications
219
70. Lee M-Y, Park CB, Dordick JS, Clark DS (2005) Metabolizing enzyme toxicology assay chip (MetaChip) for high-throughput microscale toxicity analyses. Proc Natl Acad Sci USA 102:983-987. 71. Lee M-Y, Kumar AR, Sukumaran SM, Hogg MG, Clark DS, Dordick JS (2008) Three-dimensional cellular microarray for high-throughput toxicology assays. Proc Natl Acad Sci USA 105:59-63. 72. Powers MJ, Domansky K, Kaazempur-Mofrad MR, Kalezi A, Capitano A, Upadhyaya A, Kurzawski P, Wack KE, Stolz DB, Kamm R, Griffith LG (2002) A microfabricated array bioreactor for perfused 3D liver culture. Biotechnol Bioeng 78:257-269. 73. Guguen-Guillouzo C, Clement B, Baffet G, Beaumont C, Morel-Chany E, Glaise D, Guillouzo A (1983) Maintenance and reversibility of active albumin secretion by adult-rat hepatocytes co-cultured with another liver epithelial-cell type. Exp Cell Res 143:47-54. 74. Khetani SR, Bhatia SN (2007) Microscale culture of human liver cells for drug development. Nat Biotechnol 26:120-126. 75. Bi Y, Ehirchiou D, Kilts TM, Inkson CA, Embree MC, Sonoyama W, Li L, Leet AI, Seo B-M, Zhang L, Shi S, Young MF (2007) Identification of tendon stem/progenitor cells and the role of the extracellular matrix in their niche. Nat Med 13:1219-1227. 76. Viswanathan S, Benatar T, Mileikovsky M, Lauffenburger DA, Nagy A, Zandstra PW (2003) Supplementation-dependent differences in the rates of embryonic stem cell self-renewal, differentiation, and apoptosis. Biotechnol Bioeng 84:505-517. 77. Prudhomme W, Daley GQ, Zandstra P, Lauffenburger DA (2004) Multivariate proteomic analysis of murine embryonic stem cell self-renewal versus differentiation signaling. Proc Natl Acad Sci USA 101:2900-2905. 78. Chung BG, Flanagan LA, Rhee SW, Schwartz PH, Lee AP, Monuki ES, Jeon NL (2005) Human neural stem cell growth and differentiation in a gradientgenerating microfluidic device. Lab Chip 5:401-406. 79. Abhyankar VV, Lokuta MA, Huttenlocher A, Beebe DJ (2006) Characterization of a membrane-based gradient generator for use in cell-signaling studies. Lab Chip 6:389-393. 80. Kim L, Vahey MD, Lee H-Y, Voldman J (2006) Microfluidic arrays for logarithmically perfused embryonic stem cell culture. Lab Chip 6:394-406. 81. Figallo E, Cannizzaro C, Gerecht S, Burdick JA, Langer R, Elvassore N, Vunjak-Novakovic G (2007) Micro-bioreactor array for controlling cellular microenvironments. Lab Chip 7:710-719. 82. Zhong JF, Chen Y, Marcus JS, Scherer A, Quake SR, Taylor CR, Weiner LP (2008) A microfluidic processor for gene expression profiling of single human embryonic stem cells. Lab Chip 8:68-74. 83. Hatch A, Kamholz AE, Hawkins KR, Munson MS, Schilling EA, Weigl BH, Yager P (2001) A rapid diffusion immunoassay in a T-sensor. Nat Biotechnol 19:461-465. 84. Zhao B, Moore JS, Beebe DJ (2001) Surface-directed liquid flow inside microchannels. Science 291:1023-1026.
220
Tavana, Chung, Kuo, and Takayama
85. Gu W, Zhu X, Futai N, Cho BS, Takayama S (2004) Computerized microfluidic cell culture using elastomeric channels and Braille displays. Proc Natl Acad Sci USA 101:15861-15866. 86. Futai N, Gu W, Song JW, Takayama S (2006) Handheld recirculation system and customized media for microfluidic cell culture. Lab Chip 6:149-154. 87. Wojciak-Stothard B, Ridley AJ (2003) Shear stress-induced endothelial cell polarization is mediated by Rho and Rac but not Cdc42 or PI 3-kinases. J Cell Biol 161:429-439. 88. Boo YC, Hwang J, Sykes M, Michell BJ, Kemp BE, Lum H, Jo H (2002) Shear stress stimulates phosphorylation of eNOS at Ser(635) by a protein kinase A-dependent mechanism. Am J Physiol Heart Circ Physiol 283:H1819H1828. 89. Garcia-Cardeña G, Comander J, Anderson KR, Blackman BR, Gimbrone MA, Jr. (2001) Biomechanical activation of vascular endothelium as a determinant of its functional phenotype. Proc Natl Acad Sci USA 98:4478-4485. 90. Hsai TK, Cho SK, Honda HM, Hama S, Navab M, Demer LL, Ho CM (2002) Endothelial cell dynamics under pulsating flows: Significance of high versus low shear stress slew rates (∂τ/∂t). Ann Biomed Eng 30:646-656. 91. Helmlinger G, Geiger RV, Schreck S, Nerem RM (1991) Effects of pulsatile flow on cultured vascular endothelial cell morphology. J Biomech Eng 113:123-131. 92. Lee PJ, Hung PJ, Rao VM, Lee LP (2006) Nanoliter scale microbioreactor array for quantitative cell biology. Biotech Bioeng 94:5-14. 93. Wang Z, Kim MC, Marquezcde M, Thorsen T (2007) High-density microfluidic arrays for cell cytotoxicity analysis. Lab Chip 7:740-745. 94. Lee PJ, Hung PJ, Lee LP (2007) An artificial liver sinusoid with a microfluidic endothelial-like barrier for primary hepatocyte culture. Biotech Bioeng 97:1340-1346. 95. Lee PJ, Gaige TA, Ghorashian N, Lee LP (2007) Microfluidic tissue model for live cell screening. Biotechnol Prog 23:946-951. 96. Wheeler AR, Throndset WR, Whelan RJ, Leach AM, Zare RN, Liao YH, Farrell K, Manger ID, Daridon A (2003) Microfluidic device for single-cell analysis. Anal Chem 75:3581-3586. 97. Sawano A, Takayama S, Matsuda M, Miyawaki A (2002) Lateral propagation of EGF signaling after local stimulation is dependent on receptor density. Dev Cell 3:245-257. 98. Zigmond SH, Hirsch JG (1973) Leukocyte locomotion and chemotaxis: New methods for evaluation, and demonstration of a cell-derived chemotactic factor. J Exp Med 137(2):387-410. 99. Zicha D, Dunn G, Brown AF (1981) A new direct-viewing chemotaxis chamber. J Cell Sci 99:769-775. 100. Jeon NL, Baskaran H, Dertinger SK, Whitesides GM, Van De Water L, Toner M (2002) Neutrophil chemotaxis in linear and complex gradients of interleukin-8 formed in a microfabricated device. Nat Biotechnol 20:826 - 830.
Microfluidic Systems for Cellular Applications
221
101. Lin F, Saadi W, Rhee SW, Wang SJ, Mittalb S, Jeon NL (2004) Generation of dynamic temporal and spatial concentration gradients using microfluidic devices. Lab Chip 4:164-167. 102. Wang SJ, Saadi W, Lin F, Minh-Canh Nguyen C, Jeon NL (2004) Differential effects of EGF gradient profiles on MDA-MB-231 breast cancer cell chemotaxis. Exp Cell Res 300(1):180-189. 103. Saddi W, Wang SJ, Lin F, Jeon NL (2006) A parallel-gradient microfluidic chamber for quantitative analysis of breast cancer cell chemotaxis. Biomed Microdevices 8:109-118. 104. (2003) Has the pharmaceutical blockbuster model gone bust? Bain & Company press release 105. DiMasi J (2002) The value of improving the productivity of the drug development process: faster times and better decisions. Pharmacoeconomics 20:110. 106. DiMasi J, Hansen R, Grabowski H (2003) The price of innovation: new estimates of drug development costs. J Health Econ 22:151-185. 107. Adams C, Brantner V (2006) Estimating the cost of new drug development: is it really 802 million dollars? Health Aff (Millwood) 25:420-428.
Chapter 7 Microfluidic Systems for Engineering Vascularized Tissue Constructs
Yanan Du 1,2, Donald Cropek3, Mohammad R. Kaazempur Mofrad5, Eli J. Weinberg4, Ali Khademhosseini1,2, and Jeffrey Borenstein4 1 Harvard-MIT Division of Health Sciences and Technology, Massachusetts Institute of Technology, Cambridge, MA 02139 2 Center for Biomedical Engineering, Department of Medicine, Brigham and Women’s Hospital, Harvard Medical School, Cambridge, MA 02139 3 U.S. Army Corps of Engineers, Construction Engineering Research Laboratory Champaign, IL 61822 4 Draper Laboratory, Cambridge, MA 02139 5 Department of Bioengineering, University of California at Berkeley, Berkeley, CA 94720 Correspondence should be addressed to: Ali Khademhosseini (
[email protected]) Jeffrey Borenstein (
[email protected])
Keywords: microfluidics, vascularized tissue construct, tissue engineering, hydrogel, mathematical modeling
224
Du et. al
Abstract: A major challenge in tissue engineering is the lack of proper vascularization of the fabricated tissue constructs. Microscale technologies, especially microfluidic systems, provide new opportunities to overcome the challenge of developing artificial micro-vasculatures. In this chapter, the application of microfluidic systems for generating vascularized tissue constructs by using microfluidic platforms has been outlined. Mathematical modeling has been shown as a powerful tool for optimizing the microfluidic system design parameters. Future challenges for the microfluidic-based tissue engineering constructs have also been discussed for success in clinical application.
7.1 Introduction Organ failure is one of the most serious problems faced by the healthcare industry in developed nations. Each year in the United States alone, millions of people suffer from end-stage organ failure and tissue loss, resulting in more than $400 billion in health care costs [1]. Only 10% of these patients benefit from organ transplantation, while the majority of patients perish due to the severe shortage of available organ donors [2]. To address this problem, the field of tissue engineering has emerged with the aim of generating tissues that restore, maintain, or enhance tissue function [1]. Tissue engineering is an interdisciplinary research field, which leverages both biological understanding and engineering approaches. Commonlyadopted tissue engineering approaches incorporate (i) isolated cells or cell substitutes, (ii) bio-compatible materials for cellular support and regeneration, or (iii) cell-biomaterial (i.e. scaffold) composites [3]. Transplantable cells are derived either as autografts (from the patient), allografts (from a human donor) or xenografts (from a different species). Isolated cells are then cultured on biocompatible scaffolds, which provide physical and chemical support and guide the cell growth and organization into three dimensional (3D) tissues with mimicry of living tissues in vivo (inside the body). Despite the enormous advances in tissue engineering, which have resulted in clinically viable products such as skin and cartilage, several challenges still prevent the widespread clinical application of tissue engineering products. These challenges include a number of business, regulatory and ethi-
Engineering Vascularized Tissue Constructs
225
cal issues as well as scientific barriers. These scientific issues include (1) how to acquire adequate source of cells, (2) how to engineer complex vascularized tissues that mimic the complexity of native tissue architecture and (3) how to generate tissues ex vivo (outside the body) with the biomechanical and metabolic functions that mimic normal tissues. Among all these scientific issues, one major challenge of engineering tissues in vitro (in tissue culture) is lack of a proper vascularization [4]. Development of an artificial microvasculature is critical to move tissue engineered organs into the clinics to benefit patients with end-stage organ failure. Oxygen and other nutrients can only diffuse through a short distance before being consumed (a few hundred micrometers at most). Without an intrinsic capillary network, the maximal thickness of engineered tissue is approximately 150200 µm because of oxygen diffusion limitations [5]. To date, the most successful tissue engineering applications in clinics are skin and cartilage, which have relatively low requirements for nutrients and oxygen and rely on vascularization from the host to provide permanent engraftment and mass transfer of oxygen and nutrients [6]. However, such techniques encounter difficulties when applied to thick, complex tissues, particularly those comprising the large vital organs such as liver, kidney, and heart [7]. Living tissues are ensembles of different cell types embedded in complex and well-defined geometries and within an extracellular matrix (ECM) that is unique to each tissue type. The generation of tissues requires tools to control the biological, chemical and mechanical environment experienced by cells in culture. Conventional techniques for scaffold fabrication, such as solvent casting, particulate leaching, gas forming and fiber bonding cannot be used to fabricate scaffolds with controllable pore geometry, size, and interconnectivity, that can be used to create a vasculature [8]. Several microscale technologies, especially microfluidic systems, provide new hope to overcome this challenge to build tissues with vascularized structures in a reproducible manner. Microscale technologies can achieve a resolution of 0.1 µm, which is two orders of magnitude smaller than the dimensions of the capillaries, and span of five orders of magnitude ranging from overall dimensions of tens of centimeters down to cellular dimensions of 5 µm [7]. In this chapter, we discuss the applications of microfluidic systems in tissue engineering to generate vascularized tissue constructs. The use of microfluidics for generating tissue constructs has been progressing through multiple stages of complexity over the past few years. The initial designs for the fabrication of microfluidic vascular patterns were performed in two-dimensional (2D) systems and more recently have
226
Du et. al
evolved to more biomimetic systems. Here we will provide a view of this field starting with 2D systems of microchannels with non-tissue engineering materials and progressing to more advanced systems with hydrogels and additional levels of control.
7.2 Generating 2D Vascularized Tissue Constructs Using Microfluidic Systems Most of the initial work to generate vascularized tissue constructs utilized microscale technologies (microfabrication or micro-molding) to fabricate 2D vascularized patterns (normally starts from a single channel that branches out multiple times into thinner and thinner channels) on various biomaterials (silicon, glass, or polymer). Drs. Jeffrey Borenstein and Joseph P. Vacanti were among the pioneers to build vascular tissue constructs in this manner. In their early work, photolithography was utilized to generate 2D bifurcating patterns on silicon and Pyrex, reminiscent of the branched architecture of vascular and capillary networks (Fig. 7.1A) [9]. Endothelial cells and hepatocytes (primary liver cells) were cultured and subsequently lifted from these 2D patterns as single-cell monolayers, which were subsequently shown to maintain proliferation and functionality. Single endothelial cells were aligned to form a branched network (Fig. 7.1B). In their subsequent work, soft lithographic techniques were applied to mold polydimethylsiloxane (PDMS) on silicon wafer with the impression of bifurcating vascular networks [4]. These patterned layers of PDMS were irreversibly bonded to flat PDMS layers to create enclosed network channels that were seeded with endothelial cells. To generate highly uniform flow patterns, which mimic both large-scale physiologic properties (total flow rate, i.e. ~100 ml/min in femoral artery) and small-scale phenomena (fluid velocity in the capillaries, i.e.~1 ml/min in skeletal muscle) [10], three network designs were fabricated to approximate the fluid dynamics (Fig. 7.1C). The first one provided a stepwise scaling from arteries to capillaries, but the flow resistance was much higher than physiological goals, limiting the flow under practical pressure drops to sub-optimal values. The second design provided high capillary cross section, and substantially increased the flow rate at realistic value of input pressure, but the flow was non-uniform. The third design demonstrated both uniform flow as well as reduced resistance of the network. These scaffolds have been successfully seeded with endothelial cells in PDMS channels with dimensions on the order of capillary diameters (Fig. 7.1D)
Engineering Vascularized Tissue Constructs
A
C
227
B
D
Fig. 7.1 Non-biodegradable microfluidic systems to generate 2D vascularized tissue constructs: (A) Optical image of a capillary network etched into silicon wafer (reproduced with permission from [9]); (B) H&E staining for detached monolayers of endothelial cells from the silicon capillary network (reproduced with permission from [9]); (C) Three vascularized network designs. (TEP-0: 35 x20, TEP1: 30x40, TEP-2: 25x35, Dimensions in µm) (reproduced with kind permission from Springer Science+Business Media:[4] Fig.8 ); (D) Image of confluent endothelial cells in a PDMS microfluidic vascularized structure (width of the channel: 200 µm) (reproduced with kind permission from Springer Science+Business Media:[4] Fig.10 ).
228
Du et. al
One limiting factor of microfluidic scaffolds in above two studies has been the choice of material. Microfabricated silicon and PDMS, although ubiquitous and inexpensive, are not biodegradable and have limited biocompatibility, and are therefore not suitable biomaterials for implantation. To address this limitation, a new approach for cell and tissue engineering, known as biodegradable microfluidics has been explored to fabricate potential implantable microfluidic tissue constructs [8]. King et al. have produced highly branched microfluidic scaffolds from poly (L-lactic-coglycolic acid) (PLGA) (Fig. 7.2A) [8]. Biodegradable PLGA was heated to melt on PDMS molds and then pressed with constant force between two parallel metal plates to create the vascularized patterns. The microvascularized PLGA membrane was laminated by thermal fusion bonding to construct the microfluidic systems. Although this approach of building a biodegradable microfluidic scaffold is rapid, versatile and low-cost, the potential limitations of PLGA-based biodegradable microfluidic scaffold are the cytotoxic by-products and inflammatory immune response generated during implantation [11]. Furthermore, PLGA scaffolds are brittle, hard and lack the desired mechanical elasticity of many native tissues. To address these limitations, Wang et al synthesized a novel biodegradable polymer known as poly(glycerol sebacate) (PGS) or bio-rubber [12], a tough, elastomer that is biocompatible, inexpensive, and easy to synthesize. The sebacic acid-containing polymers, which have already been approved for use in medical applications, showed reduced inflammatory response relative to PLGA [13]. Unlike PLGA, bio-rubber dissolves through a surface erosion process that is far more linear in terms of mechanical strength degradation over time, and therefore provides the resulting scaffolds more consistent mechanical integrity before degradation [12]. PGS was used to fabricate microcapillary networks by using patterned silicon micromolds as templates to fabricate PGS molds [6] (Fig. 7.2B). A patterned PGS film was then bonded with a flat film to create capillary networks that were perfused with a syringe pump at a physiological flow rate. The devices were endothelialized under flow conditions, and part of the lumens reached confluence within 14 days of culture (Fig. 7.2C). This approach may lead to tissue-engineered microvasculature that is critical in organ engineering.
Engineering Vascularized Tissue Constructs
229
Fig. 7.2 Biodegradable microfluidics to generate 2D vascularized tissue constructs: (A) A PLGA scaffold with 2D vascularized patterns made by micromolding (reproduced with With kind permission from Springer Science+Business Media:[4] Fig. 6); (B) A PGS scaffold with 2D vascularized patterns made by micromolding and (C) An expanded view of the endothelialized capillary network in PGS: Endothelial cells reached confluence at various portions of the PGS capillaries within 14 days (reproduced with permission from [6]).
Besides PLGA and PGS, other biodegradable polymeric materials have been used to build vascularized microfluidic scaffolds. Sarka et al. developed and characterized porous micropatterned poly-caprolactone (PCL) scaffolds as a functional small diameter blood vessel analog, using a novel technique that integrates soft lithography, melt molding and particulate leaching of PLGA micro/nanoparticles [14]. PCL was chosen because of its ease of processing (melting temperature 58-63oC), biocompatibility and the ability to manipulate its mechanical and degradation properties by the formation of PCL copolymers. Vascular smooth muscle cells have been aligned on these porous micropatterned PCL scaffolds. Poly(lactide-coglycolide) (PLG) has also been used to build artificial capillary networks [15]. When coated with fibronectin, endothelial cells grew to near confluence within the enclosed networks. Recently, Bettinger et al. built microfluidic devices made of the silk fibroin protein, which was from the Bombyx mori silkworm and approved by the FDA for medical applications [16]. Hepatocytes grown in the silk fibroin-based microfluidic devices exhibited similar morphology and cell functions to their counterparts cultured on other biodegradable scaffolds such as PGS.
230
Du et. al
7.3 Generating 3D Vascularized Tissue Constructs Using Microfluidic Systems In vivo, cells reside within 3D environments in close proximity to blood vessels that supply tissues with nutrients and oxygen and remove waste products and carbon dioxide. In vitro, numerous studies have identified critical features that allow 3D cultures to replicate physiology better than 2D cultures [17-20]. Thus, building 3D vascularized microfluidic scaffolds is of vital importance for the success of tissue engineering applications. Although vascularized microfluidic systems are readily constructed in 2D by photolithographic or soft-lithographic techniques, their construction in 3D remains a challenging problem. So far, the most commonly used approach is to stack and assemble 2D vascularized polymer films (usually made by micromolding techniques) into large 3D devices suitable for transplantation. King et al. developed a scalable fabrication platform for constructing highly branched, multiplayer PLGA microfluidic networks that mimicked tissue microvasculature, for large-scale tissue engineering [8]. In this approach, two or more micro-patterned PLGA films could be bonded by a pure thermal bonding process to form a monolithic, 3D and biodegradable microfluidic device (Fig. 7.3A). Bettinger et al. developed fully 3D microfabricated vascularized constructs by stacking layers of biodegradable PGS patterned films [21]. The 3D microfabricated vascular construct has the unique property of exhibiting constant maximum shear stress within each channel of the device, making it a promising construct for an artificial vascular tissue engineering scaffold. Multi-layered microfludic scaffolds have also been used for liver tissue engineering. Cheung et al. have constructed a 3D scaffold for hepatocyte culture based on a sandwich structure that places each hepatocyte compartment adjacent to a microvascular channel network in a bilayer structure [22]. Between the two chambers sits a nanoporous membrane that separates the cellular components but allows for the free exchange of oxygen, nutrients, and waste products between the compartments (Fig. 7.3B). Furthermore, Leclerc and Fujji have constructed a 3D microfluidic scaffold composed of two stacked layers of PDMS for liver cell (HepG2 cells) culture [23]. The HepG2 cells could be kept in good condition for nearly 10 days with the completely closed perfusion system.
Engineering Vascularized Tissue Constructs
231
Fig. 7.3 Microfluidics systems to generate 3D vascularized tissue constructs : (A) A multilayer PLGA microfluidic networks perfused with fluorescein dye (reproduced with permission from [8] Copyright Wiley-VCH Verlag GmbH & Co. KGaA.); (B) A 3D scaffold for liver tissue engineering based on a sandwich structure that places each hepatocyte compartment adjacent to a microvascular channel network in a bilayer structure. [22]; (C) 3D microvascular scaffold with squarespiral tower patterning fabricated by a direct-write assembly (Reprinted with permission from Macmillan Publishers Ltd: [Nature Materials][24], copyright (2003)); (D) Eight level multi-width and multi-level microvasculature network microchannels fabricated by one-step laser direct writing. Fluorescent image shows the difference in intensity levels corresponding to different channel depths ([25] Reproduced by permission of The Royal Society of Chemistry).
Building 3D vascular microfluidic structures by stacking 2D layers is a cumbersome process requiring multiple fabrication and masking steps that is difficult to scale-up. Therriault et al. demonstrated the direct fabrication of 3D microvascular networks through direct-write assembly of an organic ink [24]. Their approach was based on a 3D microvascular network of cylindrical microchannels, which could be directly assembled by robotic deposition, and then patterned to yield vertically oriented, square-spiral towers within the device (Fig. 7.3C). Sixteen-layer scaffolds were produced by robotic deposition of a paraffin-based organic ink in a layer-wise building sequence. These 3D microvascular networks will provide an enabling platform for a wide array of fluidic-based applications. In addition, Lim et al demonstrated a faster and more flexible alternative method to fabricate multiple-level microfluidic channels using a maskless laser direct micromachining [25]. Due to the inherent 2D nature of photolithography,
232
Du et. al
microfluidic channels fabricated by photolithography exhibited uniform depth, which generates non-physiological flow conditions and high flow resistances. By using a maskless laser direct micromachining, a multiwidth and multi-depth microchannel was fabricated to generate biomimetic vasculatures whose channel diameters changed according to Murray’s law (the cube of the radius of a parent vessel equals the sum of the cubes of the radii of the daughters [26]) (Fig. 7.3D). Different depths were fabricated simply by varying the average power from the laser used to micromachine the channels. The multi-depth channels, which obey Murray’s law across multiple branching generations, mimic physiological flow patterns with lower overall flow resistances and more gradual changes in the flow velocities across different generations of branching compared to channels of uniform depth.
7.4 Hydrogel-based Microfluidic Systems for Generating Vascularized Tissue Constructs Tissue engineering scaffolds made from hydrogels have aroused a great deal of interests in recent years [27]. Hydrogels are networks of hydrophilic polymers exhibiting a number of potential advantages compare to other materials such as PDMS, PGS, and PLGA, since their physical properties (i.e. mechanical strength and biodegradability) and biological properties (i.e. the biocompatibility and resemblance to the natural ECM) can be tailored to mimic tissues. Commonly-used hydrogels include natural hydrogels (i.e. collagen, hyaluronic acid, and alginate), synthetic hydrogels (i.e. poly(ethylene glycol)-diacrylate (PEGDA), and poly(vinyl alcohol) (PVA) [16]), and hybrid natural-synthetic composites [28]. Photocrosslinkable hydrogels have been used for the encapsulation of various cells [29-31], which were utilized as building scaffolds for tissue engineering [32-34]. The merger of microengineered hydrogels and microfabrication techniques for microfluidic transport has been shown of significant potential to generate 3D tissue constructs. Stroock et al. introduced a hydrogel microfluidic system within calcium alginate hydrogel [35]. They demonstrated that a high level of mass transfer could be achieved within the hydrogel microfluidic system by arraying the channels in appropriate dimensions. These results demonstrated the feasibility of using an embedded microfluidic system to control concentrations of soluble species within the 3D volume defined by a hydrogel. Recently, the same group demonstrated direct fabrication of a functional mi-
Engineering Vascularized Tissue Constructs
233
crofluidic structure within a 3D calcium alginate hydrogel scaffold for tissue engineering applications [36]. These microfluidic channels enabled an efficient exchange of solutes within the interior of the hydrogel scaffold and a quantitative control of the soluble environment experienced by the cells in their 3D environment. More than one independent vascularized network was incorporated within the microfluidic scaffolds. Each network could serve as an independent source for solutes or as a sink for others, such that concentration gradients could be maintained at steady state for both non-reactive and reactive solutes (Fig. 7.4A). This approach is promising for directing cells in the scaffolds with spatial and temporal control and growing thick sections of tissue without necrosis. Ling et al. built cell-laden microfluidic channels from hydrogels by directly encapsulating cells within the microfluidic channels [31]. Using standard soft lithographic techniques, molten agarose was molded against a SU-8 patterned silicon wafer to build microfluidic channels. Channels of different dimensions were generated and it was shown that agarose was a suitable material for performing microfluidics. Cells embedded within the microfluidic molds were well distributed and media pumped through the channels allowed the exchange of nutrients and waste products. While most cells were found to be viable upon initial device fabrication, only those cells near the microfluidic channels remained viable after 3 days, demonstrating the importance of a perfused network of microchannels for delivering nutrients and oxygen to maintain cell viability in large hydrogels. Cell-laden microfluidic hydrogels could also be scaled up by stacking the biomimetic vascular patterns to generate multi-layer vascularization in multiple discrete planes.
234
Du et. al
Fig. 7.4 Hydrogel-based microfluidics systems for generating vascularized tissue constructs: (A) Cross-sectional views of cell-seeded microfluidic scaffolds made by calcium alginate hydrogel. Dispersed cells are shown as double circles, microchannels are shown as squares. The shading represents steady-state distributions of solutes: Top, reactive solute is delivered via the channels and is consumed by cells as it diffuses into the matrix; Bottom, non-reactive solute is delivered via the two channels on the left and extracted by the channels on the right (Reprinted with permission from Macmillan Publishers Ltd: [Nature Materials][36], copyright (2007)); (B) Microfluidic systems made by collagen gel: Top, collagen gels with multi-planar network; Bottom, collagen gel with a monolayer of endothelial cells lining along the internal channels. Inset, Hoechst-stained microvascular network (reproduced with permission from [37]).
Golden et al. introduced a general procedure for the formation of microfluidic gels, with an emphasis on gels of native ECM proteins, such as type I collagen and fibrin [37]. In this approach, micro-molded meshes of gelatin were used as sacrificial materials, which were encapsulated in a second hydrogel (collagen or fibrin). The gelatin meshes were subsequently removed by heating and flushing, leaving behind interconnected channels in the second hydrogel. The channels were as narrow as 6 µm, and faithfully replicated the features in the original gelatin mesh. 50µm wide microflu-
Engineering Vascularized Tissue Constructs
235
idic networks in collagen and fibrin readily enabled delivery of macromolecules and particles into the channels and transport of macromolecules from channels into the bulk of the gels. By co-encapsulation and melting of two gelatin meshes, microfluidic gels containing two independent networks have been fabricated (Fig. 7.4B top). Microfluidic gels were also suitable as scaffolds for cell culture, which were used to culture human microvascular endothelial cells to form rudimentary endothelial networks for potential tissue engineering applications (Fig. 7.4B bottom). Growth factors have been used in combination with biomaterials to enhance the formation of microvascular networks. Richardson et al. built microvascular structures by introducing multiple growth factors (platelet derived growth factor, PDGF and Vascular endothelial growth factor, VEGF) into a PLGA scaffold in a stepwise fashion [38]. Both the spatial and temporal aspects of growth factor introduction were critical and must be orchestrated effectively in order to produce sustainable vessels. In another study, human ES (embryonic stem) cells cultured on PLLA/PLGA scaffold, when exposed to insulin-like growth factor, were able to differentiate and organize into a capillary-like endothelial network [39]. It is envisioned that the growth factor–based approaches can be incorporated within the existing microfluidic-based tissue constructs to improve vascularization.
7.5 Mathematical Modeling to Optimize the Microfluidic Systems for Generating Vascularized Tissue Constructs Establishing mathematical models of the microfluidic systems usually precedes the fabrication process. Mathematical modeling can provide powerful quantitative tools for theoretical guidance of the microfluidic design and prediction of communications between the microfluidic and biological systems in the vascularized tissue constructs, such as transport and metabolic phenomena. The earliest mathematical modeling of the vascular system could be traced back to 1926, when Murray discovered and mathematically described the relationship governing the optimum ratio between the diameters of the parent and daughter branches in vascular systems [26]. This relationship is known as Murray’s law and states that the cube of the diameter of the parent vessel must equal the sum of the cubes of the daughter vessels. For symmetric bifurcations, an important consequence of this geometric rule is that the tangential shear stress at the wall remains constant throughout the vascular network. Based on Murray’s law, multiwidth and multi-depth 3D microchannels have been fabricated by a direct
236
Du et. al
laser writing to generate biomimetic vasculatures as described in the previous section [25]. Emerson et al. generalized Murray’s law to apply to the design of constant-depth microfluidic channels and manifolds found in labon-a-chip systems [40]. A comprehensive series of computational fluid dynamics simulations considering branching networks composed of square, rectangular, and trapezoidal cross-sections were performed. These biomimetic design principles can be applied to microfluidic devices fabricated using conventional batch processing techniques without difficult multiexposure and alignment steps. Vascular geometry of the human pulmonary arterial and venous trees has also been modeled by Huang et al. [41]. The diameter-defined Strahler ordering model was used to assign branching orders, the connectivity matrix was used to describe the connection of blood vessels from one order to another, and a distinction between vessel segments and vessel elements was used to express the series parallel feature of the pulmonary vessels. Weinberg and Kaazempur-Mofrad et al. used computational fluid dynamic approaches to develop models for microcirculation [22, 42-43]. The microfluidic network was modeled as a resistor network and solved iteratively as the resistance of each vessel was altered in response to its internal pressure. The deformation of the vessel in response to an internal pressure and the resistance of the deformed vessel were calculated by numerically solving the plane-strain and Navier-Stokes equations respectively. Models for predicting hematocrit distribution within an engineered vascular network were also established [22]; similar calculations were made for pressure drop, fluid velocity, and wall shear stress within a large family of designs for tissue engineered microcirculation. Three-dimensional microfluidic constructs with uniform wall shear stress throughout the network were designed based on mathematical modeling, which could achieve more uniform endothelial cell seeding, more confluent cell coverage on the wall, and better control over cell behavior for in vitro and in vivo studies [44]. Mathematical models have been used to predict the transport and metabolic phenomenon of the microfluidic scaffolds. In the work of Choi et al., diffusivities of small and large biomolecules from the microfluidic channels into the porous hydrogel scaffold have been modelled [36]. Their results indicated that the mass transfer within the microfluidic hydrogel scaffold was diffusive but not convective. They also did computational analysis for the diffusion reaction of reactive solute (calcein-AM) and the diffusion of non-reactive solute within the bulk of microfluidic scaffolds.
Engineering Vascularized Tissue Constructs
237
7.6 Future Challenges With the rapid development of tissue engineering and microscale technologies, tremendous opportunities have been created for building vascularized tissue engineering constructs using microfluidic systems. It will continue to be one of the most important and challenging directions in the tissue engineering field. Numerous obstacles need to be overcome to build microfluidic systems for clinical applications in the future. These include: (1) how to achieve full endothelialization of the microfluidic systems with different geometries and materials; (2) how to precisely fabricate vascularized microfluidic systems with small vascularized structures that mimic the capillaries in a scalable manner; (3) how to enrich the complexity of the microfluidic tissue constructs by involving extracellular matrices, multiple cells types and controlling cell-matrix and cell-cell interactions; and finally (4) how to implant the vascularized tissue engineering constructs built ex vivo into the human body to replace lost tissue function without clogging of the blood vessels and activating the immune response.
7.7 Conclusions The merger of tissue engineering and microscale technologies opens new opportunities to build vascularized tissue engineering constructs. In this chapter, various applications of microfluidic systems in tissue engineering have been demonstrated to generate vascularized tissue constructs, which included 2D and 3D microfluidic systems (involving biodegradable and non-biodegradable polymers) and hydrogel-based microfluidic systems. Mathematical modeling has been shown as a powerful tool for optimizing the microfluidic system design. Future challenges for the microfluidicbased tissue engineering constructs have been also discussed for success in clinical applications.
References 1. Langer R, Vacanti JP (1993) Tissue engineering. Science 260:920-6. 2. Hahn MS, McHale MK, Wang E, Schmedlen RH, West JL (2007) Physiologic pulsatile flow bioreactor conditioning of poly(ethylene glycol)-based tissue engineered vascular grafts. Ann Biomed Eng 35:190-200.
238
Du et. al
3. Griffith LG, Naughton G (2002) Tissue engineering--current challenges and expanding opportunities. Science 295:1009-14. 4. Borenstein JT, Terai H, King KR, Weinberg EJ, Kaazempur-Mofrad MR, Vacanti JP. (2002) Microfabrication Technology for Vascularized Tissue Engineering. Biomedical Microdevices 4:167-175. 5. Colton CK (1995) Implantable biohybrid artificial organs. Cell Transplant 4:415-36. 6. Fidkowski C, Kaazempur-Mofrad MR, Borenstein J, Vacanti JP, Langer R, Wang Y (2005) Endothelialized microvasculature based on a biodegradable elastomer. Tissue Eng 11:302-9. 7. Borenstein JT, Terai H, King KR, Weinberg EJ, Kaazempur-Mofrad MR, Vacanti JP (2002) Microfabrication Technology for Vascularized Tissue Engineering. Biomedical Microdevices 4:167-175. 8. King KR, Wang CCJ, Kaasempur-Mofrad MR, Vacanti JP, Borenstein JT (2004) Biodegradable Microfluidics. Advanced Materials 16:2007-2012. 9. Kaihara S, Borenstein J, Koka R, Lalan S, Ochoa ER, Ravens M, Pien H, Cunningham B, Vacanti JP (2000) Silicon micromachining to tissue engineer branched vascular channels for liver fabrication. Tissue Eng 6:105-17. 10. Harrison DK, Kessler M (1989) Local hydrogen clearance as a method for the measurement of capillary blood flow. Phys Med Biol 34:1413-28. 11. Yang Y, El Haj AJ (2006) Biodegradable scaffolds--delivery systems for cell therapies. Expert Opin Biol Ther 6:485-98. 12. Wang Y, Ameer GA, Sheppard BJ, Langer R (2002) A tough biodegradable elastomer. Nat Biotechnol 20:602-6. 13. Sundback CA, Shyu JY, Wang Y, Faquin WC, Langer RS, Vacanti JP, Hadlock TA (2005) Biocompatibility analysis of poly(glycerol sebacate) as a nerve guide material. Biomaterials 26:5454-64. 14. Sarkar S, Lee GY, Wong JY, Desai TA (2006) Development and characterization of a porous micro-patterned scaffold for vascular tissue engineering applications. Biomaterials 27:4775-82. 15. King KR, Terai, H., Wang, C.C., Vacanti, J.P., and, Borenstein JT (2001) Microfluidics for tissue engineering microvasculature: Endothelial cell culture. In: Proceedings of the Fifth International Conference on Miniaturized Chemical and Biochemical Analysis Systems, Monterey, CA, pp 247. 16. Bettinger CJ, Cyr KM, Matsumoto A, Langer R, Borenstein JT, Kaplan DL (2007) Silk Fibroin. Microfluidic Devices Advanced Materials 19:2847-2850. 17. Du Y, Chia SM, Han R, Chang S, Tang H, Yu H (2006) 3D hepatocyte monolayer on hybrid RGD/galactose substratum. Biomaterials 27:5669-80. 18. Matsumoto T, Yung YC, Fischbach C, Kong HJ, Nakaoka R, Mooney DJ (2007) Mechanical strain regulates endothelial cell patterning in vitro. Tissue Eng 13:207-17. 19. Toh YC, Zhang C, Zhang J, Khong YM, Chang S, Samper VD, van Noort D, Hutmacher DW, Yu H (2007) A novel 3D mammalian cell perfusion-culture system in microfluidic channels. Lab Chip 7:302-9.
Engineering Vascularized Tissue Constructs
239
20. Sudo R, Mitaka T, Ikeda M, Tanishita K (2005) Reconstruction of 3D stackedup structures by rat small hepatocytes on microporous membranes. Faseb J 19:1695-7. 21. Bettinger CJ, Weinberg EJ, Kulig KM, Vacanti JP, Wang JT, Borenstein JT, Langer R (2006) Three-Dimensional Microfluidic Tissue-Engineering Scaffolds Using a Flexible Biodegradable Polymer. Advanced Materials 18:165169. 22. Borenstein JT, Weinberg EJ, Orrick BK, Sundback C, Kaazempur-Mofrad MR, Vacanti JP (2007) Microfabrication of three-dimensional engineered scaffolds. Tissue Eng 13:1837-44. 23. Leclerc E, Sakai Y, T F (2003) Cell culture in 3-Dimensional microfluidic structure of PDMS. Biomedical Microdevices 5:109. 24. Therriault D, White SR, Lewis JA (2003) Chaotic mixing in three-dimensional microvascular networks fabricated by direct-write assembly. Nat Mater 2:26571. 25. Lim D, Kamotani Y, Cho B, Mazumder J, Takayama S (2003) Fabrication of microfluidic mixers and artificial vasculatures using a high-brightness diodepumped Nd:YAG laser direct write method. Lab Chip 3:318-23. 26. Murray CD (1926) The Physiological Principle of Minimum Work: I. The Vascular System and the Cost of Blood Volume. Proc Natl Acad Sci U S A 12:207-14. 27. Nguyen KT, West JL (2002) Photopolymerizable hydrogels for tissue engineering applications. Biomaterials 23:4307-14. 28. Flores-Ramirez N, Elizalde-Pena EA, Vasquez-Garcia SR, GonzalezHernandez J, Martinez-Ruvalcaba A, Sanchez IC, Luna-Barcenas G, Gupta RB (2005) Characterization and degradation of functionalized chitosan with glycidyl methacrylate. J Biomater Sci Polym Ed 16:473-88. 29. Yeh J, Ling Y, Karp JM, Gantz J, Chandawarkar A, Eng G, Blumling J, 3rd, Langer R, Khademhosseini A (2006) Micromolding of shape-controlled, harvestable cell-laden hydrogels. Biomaterials 27:5391-8. 30. Elisseeff J, McIntosh W, Anseth K, Riley S, Ragan P, Langer R (2000) Photoencapsulation of chondrocytes in poly(ethylene oxide)-based semiinterpenetrating networks. J Biomed Mater Res 51:164-71. 31. Ling Y, Rubin J, Deng Y, Huang C, Demirci U, Karp JM, Khademhosseini A (2007) A cell-laden microfluidic hydrogel. Lab Chip 7:756-62. 32. Jager M, Degistirici O, Knipper A, Fischer J, Sager M, Krauspe R (2007) Bone Healing and Migration of Cord Blood derived Stem Cells into a Critical Size Femoral Defect after Xenotransplantation. J Bone Miner Res 22:1224. 33. Hahn MS, Teply BA, Stevens MM, Zeitels SM, Langer R (2006) Collagen composite hydrogels for vocal fold lamina propria restoration. Biomaterials 27:1104-9. 34. Fedorovich NE, Alblas J, de Wijn JR, Hennink WE, Verbout AJ, Dhert WJ (2007) Hydrogels as extracellular matrices for skeletal tissue engineering: state-of-the-art and novel application in organ printing. Tissue Eng 13:190525.
240
Du et. al
35. Cabodi M, Choi NW, Gleghorn JP, Lee CS, Bonassar LJ, Stroock AD (2005) A microfluidic biomaterial. J Am Chem Soc 127:13788-9. 36. Choi NW, Cabodi M, Held B, Gleghorn JP, Bonassar LJ, Stroock AD (2007) Microfluidic scaffolds for tissue engineering. Nat Mater 6:908-15. 37. Golden AP, Tien J (2007) Fabrication of microfluidic hydrogels using molded gelatin as a sacrificial element. Lab on a Chip 7:720-725. 38. Richardson TP, Peters MC, Ennett AB, Mooney DJ (2001) Polymeric system for dual growth factor delivery. Nat Biotechnol 19:1029-34. 39. Levenberg S, Huang NF, Lavik E, Rogers AB, Itskovitz-Eldor J, Langer R (2003) Differentiation of human embryonic stem cells on three-dimensional polymer scaffolds. Proc Natl Acad Sci U S A 100:12741-6. 40. Emerson DR, Cieslicki K, Gu X, Barber RW (2006) Biomimetic design of microfluidic manifolds based on a generalised Murray's law. Lab Chip 6:447-54. 41. Huang W, Yen RT, McLaurine M, Bledsoe G (1996) Morphometry of the human pulmonary vasculature. J Appl Physiol 81:2123-33. 42. Weinberg EJ, Kaazempur-Mofrada MR, Borenstein JT(2001) Numerical model of flow in distensible microfluidic network. Computational Fluid and solid Mechanics 864. 43. Kaazempur-Mofrada MR, Vicanti JP, Kamm RD (2001) Computational modeling of blood flow and rheology in fractal microvascular networks. Computational Fluid and solid Mechanics 864. 44. Weinberg EJ, Bornstein JT, Kaazempur-Mofrada MR, Orrick B, and Vacanti JP (2004) Design and Fabrication of a Constant Shear Microfluidic Network for Tissue Engineering. In: Mat. Res. Soc. Symp. Proc. Vol. 820
Chapter 8 High Throughput Screening Using Microfluidics
John P. Puccinelli and David J. Beebe Department of Biomedical Engineering, University of Wisconsin–Madison Madison, WI 53706 Correspondence should be addressed to: David Beebe (
[email protected])
Keywords: high throughput, microfluidic, cell culture, electrophoresis, PCR, drug screening
Abstract Microfluidic systems are well suited to create platforms for high throughput screening as they offer the possibility to reduce reagent volumes and increase functionality without increasing the experimental footprint. This chapter explores the current literature for high throughput microfluidic platforms in the areas of cellular and biochemical assays as well as drug screening, presenting the benefits and limitations of each system. Four main microfluidic methodologies for obtaining high throughput functionality emerge and are compared: isolated channel networks that do not require external pressure sources, flow or perfusion based fluidic devices, those that utilize arrays of valves and actuators and droplet based systems such as electrowetting on a dielectric (EWOD). In addition the role and interaction of users and developers of these systems i.e research labs, academic
242
Puccinelli and Beebe
screening facilities and industry are presented in the context of the current trends and outlook for high throughput microfluidics.
8.1 Introduction From the time high throughput screening (HTS) began in the 1980’s [1] its global market has increased to an estimated $7.6 billion in 2005 with over 400 companies involved in HTS services or instruments [2]. As this market persists, new technologies continue to emerge for applications in cell biology, molecular biology, biochemistry, chemistry, proteomics, diagnostics and pharmacology. Using robotic liquid handlers, screening over 100,000 compounds per day is possible. Traditionally HTS is performed in well plates (Table 8.1). A demand for increased throughput has led to more, smaller wells per plate reducing the reagent volumes required while increasing the surface area-to-volume ratio (SAV) of the wells. The continued miniaturization of wells has exacerbated evaporation and edge effects, in turn causing increased background and reduced signal-to-noise ratios [3, 4]. Table 8.1 Well volumes and plates. Well Plate Format Average volume/well (µL) Average cells per well
average cell number per well for standard well 96 [5] 200 40,000
384 [5] 50 5,000
1536 [5] 8 4,000
3456 [4] 2 700
Microfluidic systems have been applied to similar areas as HTS involving the integration of engineering and physics with biochemistry and biology. To do so, scientists have taken advantages of the increasing role played by laminar flow, diffusion and surface tension at the microscale [6, 7]. The typical geometry of microchannels provides an increased SAV ratio with reduced edge effects as compared to open-well formats. Microfluidics has been applied to a wide range of fields including biology and biochemistry [8], cell biology [3, 9, 10], neurobiology [11], chemistry [12], diagnostics [13], drug discovery [14] and the global health market [15]. As microfluidics is ported into HTS array formats, various fluidic manipulation techniques have been developed to provide increased fluidic control, such as valving, multiplexing, pumping, and mixing [16]. Microfluidics by definition parallels the HTS field as both strive to reduce reagent volumes and experimental footprint (size of the assay platform) while increasing scalability. As throughput continues to increase in titer
High Throughput Screening Using Microfluidics
243
plate formats from 96 to 384 to 1536 and to 3456 well plates, the term “high” with respect to throughput is still relative in the microfluidic community. With the advent of microfluidic high throughput screening (µfHTS), a continuum of throughput arises ranging from flasks and well plates to microfluidic systems such as arrays of individual channel networks (ICNs), flow and perfusion, valve-based, and droplet-based systems (Fig. 8.1). Novel microfluidic platforms can provide decreased sample size, shorter analysis times, increased sensitivity and functionality, but to date have not been competitive with traditional HTS in terms of throughput. There are two general approaches to increasing throughput – integrating steps serially on a device or side-by-side parallelization. Integrated microfluidic systems will potentially benefit the life sciences by reducing process times via combining functionalities in serial operations. Such systems include those that merge Deoxyribonucleic acid (DNA) purification, polymerase chain reaction (PCR), product separation and analysis on a single chip [17] or those that combine cell culture with analyte detection [18]. While serial operation systems may find use in areas such as point of care diagnostics, they will not be covered in this chapter. Parallel systems perform numerous operations simultaneously side-by-side and will be the focus of this chapter. These can include parallel processing of serial operations as well. Here we will define high throughput as more than 96 parallel operations. The term moderate throughput will be used for less than 96 operations. Low throughput approaches of just a few operations will not be discussed even if there is “potential” to scale upward. This chapter will focus on the use and prospective benefits of microfluidics to perform biological assays in high throughput; specifically it will focus on cell-based and biochemical assays as well as drug screening applications. As this is a rapidly changing field, this chapter provides a current snapshot of work in this area and discusses the fundamental issues and challenges that need to be considered as microfluidics attempts to become high throughput in a practical sense.
244
Puccinelli and Beebe
Fig. 8.1 Graphic illustrating the continuum of traditional HTS formats (flasks, 96, 384, 1536, 3456 well plates) as compared to four types of microfluidic methods: isolated channel networks (ICNs) such as the micro conduit array (MCA) [19], flow or perfusion-based fluidic devices [20], those that utilize arrays of valves [21] and droplet based or EWOD [22]. Each was assigned scores for ease of use in the HTS community, isolation of conditions or wells from neighboring conditions or wells and minimal dead volume of reagents (including volumes in connections). Microfluidics has the potential to span the range of HTS formats in terms of volumes and throughput as well as improve performance.
8.2 Cell-Based Assays Microfluidic systems are potentially well-suited for cell biology. The ability to create defined microenvironments [9, 23-25] through diffusion dominant systems [26, 27], generate complex gradients for cell migration [2831] and manipulate the fluidics for applications such as sorting [32] show promise. As cells in vivo exist within ~100 µm of the nearest capillary, microfluidic networks can mimic this native architecture with the ability to create defined fluidic paths for perfusion type cultures [9, 33, 34]. However, current materials used in constructing microfluidic networks – such as poly(dimethylsiloxane) (PDMS) – have been shown to absorb small hydrophobic molecules [35, 36], while other materials non-specifically ad-
High Throughput Screening Using Microfluidics
245
sorb various proteins [37, 38]. In addition, the drastic reduction in the SAV ratio may lead to changes in the cell microenvironment due to nutrient depletion, waste accumulation and evaporative osmolarity shifts – all of which can ultimately affect cell phenotype [26, 39]. As a result, the design and use of microfluidic systems for cell-based assays require thoughtful consideration of these issues including appropriate biological controls and validation. Utilizing the microfluidic scale and control, cells have been integrated into lab-on-a-chip technologies for numerous applications including cell culture, treatment, sorting, lysis, separation, and analysis [3]. 8.2.1 High throughput cell culture
Culturing cells in the laboratory has advanced our understanding of biology, development, and disease since the early 1900s while becoming more wide spread in the 1950s. In vitro cell culture has also provided a means to screen drug candidates prior to animal testing. As the demand for increased throughput and improved culture models continues to rise, microfluidics offers a promising direction for potential advances as it has already been applied to HTS for culture optimization, gene expression analysis, 3D culture and co-culture. In addition, various techniques have been used for these applications such as fluidically connected microwells, upstream gradient generators, elaborate valving and pumping mechanisms, as well as isolated channel networks for use with robotic pipettors. Microfluidic devices are typically closed, self-contained systems, thus making it difficult to perform long term maintenance of cell cultures, specifically cell passage. This has limited the culture time to around 3 days (cell type dependent) when cell confluency occurs. Through a 10 x 10 continuous perfusion cell culture array, the functionality of traditional flask type culture has been realized in a format the size of a well in a 1536 well plate (Fig. 8.2) [20]. In addition, a gradient generator (similar to that presented by others [40]) was placed upstream of the array providing 10 unique concentrations to the system with 10 replicates each; or, by utilizing an orthogonal perfusion inlet 100 conditions can be created. Cell culture under perfusion for 16 days was demonstrated, during which time all cell chambers were passaged simultaneously and then the cells were allowed to plate again within the same device. As this design showed a cell growth dependence on the perfusion rate, they later developed a device with a slightly different design. Here perfusion occurred through highly resistive elements to decrease the flows and shear over the cells while allowing for media/stimuli replenishment [27]. The new design also incorpo-
246
Puccinelli and Beebe
rated c-shaped wells that provided increased cell number uniformity between wells compared to the pervious design. A significant limitation is the need for continuous perfusion; consuming regents, washing away the effect of cell-cell communication or possibly leading to cross contamination between conditions. Other design considerations for perfusion-based cultures have been reviewed [41].
Fig. 8.2 Flow-based systems operate via the continual perfusion of media and/or gradients for concentration based cytotoxicity testing. A) An overview of the device containing both functionalities: a gradient generator (upper left) and the perfusion inlet (lower left). B) The outcome of the 10 x 10 array post-gradient generation where the insert depicts the details of each well containing small perfusion channels [20] – Copyright © (2005 Biotechnology and Bioengineering); Reprinted with permission of Wiley-Liss, Inc., a subsidiary of John Wiley & Sons, Inc.
The need for optimizing culture conditions, i.e. medium composition, medium replenishment, surface treatment and seeding density, lends itself to HTS. Quake’s group developed an automated individually addressable 96 well microfluidic array with a well volume of 60 nL per well [21]. The system, once programmed, can be left unattended for cell culture and imaging. Sixteen inputs, inline with a multiplexer, create unique fluid mixtures for each well. In this system they studied the time dependent effect of osteogenic differentiation media on the motility of primary human mesenchymal stem cells. The complexity of the device, however, is a drawback having numerous tubing inputs/outputs leading to increased dead volume and the need for high precision control equipment. Despite the low volume chambers, the device needs ~2 mL of medium/day, of which 98% is required for fluidic path washing. In addition, failure anywhere in the valv-
High Throughput Screening Using Microfluidics
247
ing schemes may lead to undesired or unnoticed cellular cross talk conditions as compared to a system with fluidically isolated chambers. Beyond performing basic culture in microfluidic arrays, there is an interest in the ability to analyze cellular responses such as gene expression. By incorporating green fluorescent protein (GFP) reporters in the cell’s DNA, dynamic gene expression can be studied in high throughput. This was done in an array of 256 live cell microfluidic bioreactors, each with a 10 nL volume (Fig. 8.3) [42]. The array was separated by two valving networks – one to allow seeding into the rows and the other to allow stimulus flow down the columns. This divided the array into 64 unique experimental condition zones with 4 replicates in each zone. In this way they were able to take advantages of microfluidics to provide rapid loading of 8 different reporter cell constructs with constant perfusion of media or 8 different stimuli across each construct. In the system they studied hepatocyte inflammation in response to bacterial endotoxin, hormones, cytokines and combinations thereof, dynamically imaging every 90 minutes over 18 hours ending with ~5000 data points. Despite drastically reduced assay times and increased throughput over traditional gene expression analysis, the system remains complex. To simplify the system, they actuated all the valves with two syringes, one syringe pump and serial loading. To simplify and add temporal control, the group later looked at similar responses in an array containing two merging inputs that when they meet, are separated by laminar flow [43]. Depending on the input flow ratio, only a particular region of the array is treated by the stimulus. They termed this “flow encoded switching.” By coupling this with upstream gradient generators specific regions of the array can be treated with specific gradients at various times. The array can also be arranged in series and parallel to obtain the desired effect. Compared to manual fluid switching, which requires washes and many fluidic manipulations, this perfusion based system provides the ability to study the long term effect on gene expression of various stimuli, stimuli pulse duration, frequency and dose. However, currently it requires up to 60 seconds to completely switch on/off a stimulus in a region.
248
Puccinelli and Beebe
Fig. 8.3 Valve operated systems typically contain loading channels and valve control channels; a) shows an overview of the 16x16 array and b,c) demonstrates valve operation via pressure through the valve control channels. d,e) Depicts the operation of the device where first, cell seeding valves are opened; second, 16 different reporter cell lines flow across the channels; third, closing all valves permit the cells to plate; fourth, opening the orthogonal stimuli valves allow for fifth, treating the cells with 16 different stimuli ending with 4 replicates each for live cell gene expression analysis [42] – Reproduced by permission of The Royal Society of Chemistry.
It has been shown that the cellular microenvironment, i.e. the interaction between cells and the extracellular matrix both mechanically and through receptor attachment/transduction, affect the cellular phenotype and genotype [44]. This had piqued interest in culture methods that more faithfully recapitulate in vivo conditions such as 3D constructs and co-cultures. However, the properties (i.e. viscosity and thermal setting) of various 3D constructs (agarose, gelatin, collagens and Matrigel™) pose obstacles for use in microsystems. Viscous materials that polymerize can clog channels, coat surfaces, make cell removal difficult, and have reduced flow in more resistive microchannels. Such materials require temperature control and separate fluidic paths to reduce these problems. For example, gelatin has been used to fabricate channels, allowing cells to migrate into the channel bottom directly [23, 45]. With the inherent difficulties of 3D culture in microchannels, work in this area has, in general, been limited to moderate throughput. An array etched in Si with PDMS channels has been used for 3D collagen cell culture and drug screening in a 4x5 array [46]. In the ar-
High Throughput Screening Using Microfluidics
249
ray, the chemosensitivity of breast cancer cells to cancer drugs in 2D and in 3D was compared. The group later utilized a similar design to test the effect of four stimuli on hepatocyte and breast cancer spheroids for two weeks [47]. These systems, however, are still manually loaded, limiting throughput. Automated 3D loading was demonstrated within a 30microbioreactor perfusion culture [48]. The 6-layer PDMS device contained s-shaped pneumatic micropumps that loaded the cell/agarose solution as well as provided pressure for perfusion medium. More work is needed in this area to increase 3D culture throughput such as creating arrays of channels using the matrix as a construct, isolating 3D cultures to a specific region of a channel network to decrease the diffusion distances for nutrients or by using a different method to load the constructs to prevent clogging. Not only does the extracellular matrix play a pivotal role in cell behavior, but the interaction of multiple cell types during development and tumorogenesis often leads to altered morphology, expression and differentiation as compared to monoculture [49-51]. The ability to study these interactions in high throughput could speed development of new therapeutics. Utilizing various microfabrication and patterning techniques, Khademhosseini et al. created a co-culture by first constructing microwells in PDMS and then placing an array of PDMS channels on top. Different cell types were then introduced by flow into each channel and allowed to settle into the wells. The channels were finally removed and replaced with orthogonal channels creating an array of co-culture channels [52]. They also demonstrated the use of an upstream gradient generator to apply different conditions to each co-culture. Flow though the system, however, eliminates secreted factors and thereby negates paracrine interactions and the co-culture effect. The array was also only limited to 5x5 to minimize complexity of connections to the outside world. In addition, PDMS delamination can be a potential problem when laying down channels post protein adsorption, which can modify the surface properties. While the new channels do bond to the modified surface, it is not as strong as bonding to a virgin surface resulting in delamination under higher pressures. As this technique is a step to HTS co-culture, more work is needed to advance this area. A different approach consisting of isolated channel networks (ICNs) requires no tubing, valving or external pressure sources. This approach operates by surface tension based pumping from the internal pressure of a droplet termed “passive pumping” to manipulate fluid in separate individual channels [53]. Passive pumping operates via surface tension differ-
250
Puccinelli and Beebe
ences between droplets at two entrances of a filled microchannel and is thus governed by the pressure gradient between these drops, as determined by the Young-LaPlace equation (Fig. 8.4A). Pumping is achieved by placing a larger droplet (negligible internal pressure, ~5-10 μL) on one port of the filled channel and subsequently placing a smaller droplet (high internal pressure, ~2-5 μL) on a second port: pumping then occurs point-to-point from the smaller drop to the larger drop. Since pumping is point-to-point, any number of channels, channel geometries or channel entrances/exits can be operated robustly and in a defined manner with a single pipettor, a multichannel pipettor or an automated robotic liquid handler for fluid addition/removal depending on the throughput desired [19]. The channels can be tightly arrayed without cross talk between them in various formats to easily interface with current liquid handling technology, these devices have previously been termed microconduit array (MCA). The MCA has been characterized with 192 individual cell culture channels having a volume ~1 µL each and a 384 well plate format for access to the ports [54]. Density up to 1536 well plate format (768 channels) has also been demonstrated with cell culture. MCAs have been used to seed cells, replace media, treat, fix and stain cells. A direct comparison of a cell line and primary cells was tested with various culture conditions in “wiggle channels” that occupy the space of a 96 well plate (Fig. 8.4B) for assay readout on a plate reader [55]. The study examined differences in culture conditions (cell number with uniform seeding density and media components). It also reduced the total number of primary cells used for this assay 20 times compared to 24 well plates typically used for this assay. As a result, samples could be obtained from a single animal reducing variability between conditions. Various other channel geometries have been designed for applications in co-culture (Fig. 8.4C), penta-culture, cell migration and wound healing [56-58]. Using passive pumping has also proven to be more effective in consuming less volume for channel washing compared to traditional well plates [59]. Washing flow rates and slow flows for perfusion were modeled with this system [60]. In addition, the individual channel structure of the MCA allows for increased throughput without any cross contamination between conditions. Also, passive pumping has been utilized to generate digital logic gates, offering the opportunity to design fully autonomous, timed and automated assays simply by placing one or more droplets at one point in time and allowing the assay to proceed [61]. The system currently is sensitive to evaporation, as are many small volume systems, which can adversely affect cell culture conditions. Various methods have been employed to reduce, compensate, and model these effects, the most effective of which is to maintain the MCA in a humidified environment [62].
High Throughput Screening Using Microfluidics
251
Fig. 8.4. Isolated channel networks require no tubing connections or external pressure sources. A) A schematic of passive pumping in a single straight channel with two ports. A large drop is placed on one port. When a small drop is pipetted onto the other port, surface tension based pumping occurs where the small drop collapses and is pumped to the large drop replacing the contents of the channel [56]. B) By fitting the channels into a 96 well plate format assay readouts can be performed on a plate reader. Each row of channels has a different volume to study the effect of cell number but not surface density with various media compositions [63]. C) Co-culture can also be performed in high throughput with ICNs. Here two wells are connected to one common output. Using a multi-channel pipettor two cells types can be seeded into the wells using passive pumping. The media in the 6x10 array can then be replaced with stimuli containing media for unique cocultures [56].
The fluidic control of microsystems provides an avenue for high throughput cell culture whether for culture condition optimization, drug screening, or cellular scale explorations i.e. with 3D culture, co-culture or migration. Within these various applications, three main microfluidic methods emerge: flow based systems for perfusion of nutrients or gradients, valve controlled fluidic conditions and isolated channel networks. The latter requires no external pressure sources and can easily integrate into current HTS liquid handlers whereas the other two methods require outside pressure sources such as syringe pumps and other controllers. In all systems, the microenvironment is an important consideration. Due to the small SAV ratio in microfluidics, diffusion dominates under no flow conditions which may contribute to a build up of secreted factors more on the scale of the in vivo environment. Whereas in well plates or perfusion systems convective mixing, dilution or removal leads to a different environment [64].
252
Puccinelli and Beebe
8.2.2 Cell sorting for high throughput applications
Following the extraction of primary cells or at the end point of cell cultures, whether in macro culture or micro, often there are specific populations of cells of interest. These require identification and/or separation from the total population based on various properties of the cell such as cell surface markers, morphology, size or magnetic moment [65]. A common method of sorting is fluorescence activated cell sorting (FACS) which may require multiple sorts to obtain a pure population. FACS is widely used but can be physically harsh on the cells lowering the viability of the sorted cells. The ideal method would provide purity while maintaining a high percentage of viable cells. Conventional FACS machines can sort 10,000 cells per second, where microfluidic FACS systems are currently around two orders of magnitude slower, but often provide a higher percentage of viable cells [14]. Therefore, there is a trade off between throughput and purity/quality. A higher purity and quality is especially important for rare cell applications and separation of primary cells. Microfluidics has the potential to provide a less detrimental environment based on scale thereby reducing shear stresses. In addition smaller starting numbers and volumes can be sorted more effectively and parallel operation is more amenable. Macroscopic systems require 105-106 cells/mL where microscale devices can sort down to 102-103 cells/mL which is also advantageous for rare cells applications. Through integrated systems, microfluidics can incorporate down stream analysis in a single chip. MicroFACS systems have been developed and by combining a series of valves and peristaltic based pumps with a CCD camera the sorting of as few as 1-10 rare cells from a 107 cell/mL suspension in a 1 µL drop was automated [66]. This system, being one of the faster microFACS at 1500 cells/s, is still an order of magnitude slower than traditional FACS, yet was able to sort a higher enriched population of mammalian cells with less determent to the cell. A similar method developed by the Quake group incorporated micropumps and micro valves to sort cells from small volumes with a maximum of only 44 cells/s and enrichment up to 89 percent [67]. Another method of cell sorting employs optical trapping/tweezers, which uses a focused laser to move dielectric objects. This method offers the ability to visually or automatically identify cells/particles of interest and move them to a specific location. The Voldman group fabricated a 30-microwell device under a fluidic layer where each well holds a few cells [68]. Using an optical tweezers-like technology, termed “optical firehose,” a lowdivergence beam is able to manipulate a cell in 2D, allowing a single cell
High Throughput Screening Using Microfluidics
253
to be raised into the fluidic path and collected from the device. The result is a well or a collection tube with the desired cells remaining where further experiments can be performed. While this method is slow requiring <10 seconds per well, the system can be fully automated based on cell fluorescence and integrated into standard fluorescence microscopy and scaled up. The system also provides a pure population with a minimal risk to the cells making it appealing for rare cell applications. Using similar principles, another group used an optical lattice to sort particles of various size or refractive index from flow in a microchannel [69]. The particles or cells of interest are deflected into an upper channel while the others remain in the laminar stream lines obtaining near 100% purity at ~25 particles/s. Using an optical switch in a similar Y-shaped channel, up to 100 cells/s were sorted and a maximum purity of 98% was obtained [70]. Cells can be sorted at high efficiency by applying a non-uniform electric field on a dielectric particle in dielectrophoresis (DEP) [71, 72]. If labeled, the electric field would divert the cells into the collection channel under flow conditions; otherwise they would default to the waste channel. With this method 10,000 cells/s were sorted reaching a 250 fold enrichment and with a second round achieved 65% purity. While this method may have higher throughput, purity is scarified as compared to optical methods. A non-labeling method for cell sorting utilizes protein-functionalized pores that interact with a specific cell surface marker [73]. The electrical resistance of the pore changes when cells containing the desired surface marker pass through the pore and the speed of these cells is drastically reduced. This method can be used to determine the relative numbers of cell types or cells that have differentiated in a culture with precision. Microfluidic cell sorters while not currently as fast as FACS, tend to be less detrimental to the cells. In addition non-labeling methods have been developed such as optical properties or functionalized pores to sort cells. Following identification of a cell type of interest analysis can be done on chip as well. Further microfluidic cell based technology can also be found in Chapter 5 of this book (Microfluidic Systems for Cellular Applications).
8.3 Biochemical Assays The microscale offers advantages in performing numerous biochemical assays in parallel over conventional well plates. Most notable is the ability to create defined fluidic paths and environments such as temperature changes
254
Puccinelli and Beebe
for performing PCR or the addition of electrodes to perform electrophoretic separations of peptides, proteins, nucleic acids, and other biomolecules. In addition, biochemical analysis can include detection and immunoassay readouts utilizing the small sample volumes of microfluidic systems. 8.3.1 PCR
Genome sequencing, gene analysis, and genetic disease diagnosis all require large quantities of a DNA or mRNA sequence(s) which can be amplified using PCR or RT-PCR (reverse transcriptase) respectively. However, PCR requires roughly 20-40 cycles involving three or more temperature changes for denaturing, primer annealing and extension/elongation typically taking 5-15 minutes per cycle using traditional methods. The small SAV ratio in microsystems supports fast thermal cycling and heat energy transfer and, in addition, allows less than nanoliter sample sizes to be used [74]. The advantages have prompted the use of microfluidics for PCR incorporating many valving, pumping, and mixing functionalities [75]. Using parallel microfluidics and pneumatic actuation to shuttle samples over three temperature zones, 100 samples were amplified in less than 10 minutes [76]. Microfluidics offers similar advantages for synthesizing the PCR primers, small oligonucleotides, or genes. Gene synthesis in a 3698 chambered (each chamber with a picoliter volume) device (Fig. 8.5A) with 10 µL total volume created high precision 30-45mer oligonucleotides that then were assembled into gene products [77]. Multiplexing PCR offers the ability to run multiple sets of primers on the same sample of DNA. This reduces processing time and reagent consumption and provides multiple genes of interest which can be applied to diagnostics. Comparing genes can reveal mutations and relative transcript number for expression analysis. For this the amplified products require two separations. These functionalities were integrated into one device containing multiple reaction chambers and separation channels [78]. Using 2D electrophoresis separation of multiplex PCR products first separated them by size and then subsequently each product can be separated by single nucleotide polymorphisms (SNP) using temperature gradient gel electrophoresis (TGGE) in a single parallel arrangement microfluidic device in 5 minutes [79].
High Throughput Screening Using Microfluidics
255
8.3.2 Electrophoresis
Separations via electrophoresis have advanced biology and genomics through identifying activation cascades, genetic variation, SNP and for genetic diagnostics. Many traditional methods require extensive procedures for both electrophoretic separation and subsequent detection where the properties of the microfluidic system such as optical properties, fluidic control, and ability to incorporate electrodes enable both to be completed on-chip increasing throughput. This has led to development of micro capillary array electrophoresis (µCAE). Early work by the Mathies group demonstrated the use of µCAE in a 96 sample array that was used to separate dsDNA restriction enzyme fragments in identifying a genetic disease mutation that leads to the addition of a restriction site [80]. They report an average run and analysis (via laser scanning the array) time of 8 minutes with greater than 99.9% confidence in genotyping, increasing throughput 50100 fold over tradition gel electrophoresis. Sample loading requires the most time, however, they utilized the traditional 96 well plate spacing for multi-channel fluid handling systems. More recently they have developed a radial 384 µCAE device (Fig. 8.6) for genetic analysis where sequencing and analysis of genetic variation for 384 individuals was run in 7 minutes with 98.7% accuracy [81]. Other publications in microfluidic applications in electrophoresis have been reviewed [82] as well as microfluidic separations of chiral compounds [83]. One of the major limitations and time constraints of these systems as well as traditional systems is sample loading. 8.3.3 Others
Other various biochemical assays have been performed in high throughput utilizing the properties of microfluidic systems. An integrated system combining cell culture with µCAE and time-lapse immunoassays measured the secretion of insulin from four parallel culture regions every 6.25s resulting in 1450 immunoassays in 40 minutes [84]. Quake’s group developed a device to measure the binding energy for low affinity proteinprotein and protein-DNA binding [85]. The device consisted of 2400 chambers and was controlled by 7233 valves (Fig. 8.5B). Array based formats of immobilized biomolecules have also been integrated into microfluidic systems [86]. Takeuchi’s group has created arrayed microfluidic methods to form lipid bilayers for studying integral membrane proteins such as ion channels in which they successfully measured ion currents for a 5x5 array and have more recently scaled to over 96 wells [87].
256
Puccinelli and Beebe
Fig. 8.5 A) Gene synthesis was performed in 3698 picoliter volume wells where oligonucleotides were only added where the light beam hits [77] – Reprinted by permission of Oxford University Press. B) A device to measure binding energy of transcription factors that contains 2400 chambers and 7233 valves [85] – Reprinted with permission from AAAS.
High Throughput Screening Using Microfluidics
257
Fig. 8.6 A 384 lane micro capillary array electrophoresis radial array with the detection region centered in the middle for simplified readout. Reprinted with permission from [81] – Copyright © 2002 American Chemical Society.
Biochemical assays have also been completed in droplets. Electrowetting on a dielectric (EWOD) allows precise control of nanoliter size droplets by changing the hydrophobicity of a surface, thereby eliminating the need for fluidic manipulation tools such as pumps and valves. Other droplet based methods use surface energy differences to create droplets in immiscible fluids. Droplet based fluidics is highly applicable to biochemical assays providing isolation and/or defined mixing of conditions as well as drastically reduced volumes. This technology allows microscale control to perform complete assays within drops [88], diagnostics using physiological fluids i.e. whole blood, serum, plasma and urine [89], separation of particles in drops [90] and PCR [91] and this method has recently been reviewed [22, 92]. It, however, has only minimally been applied to cell biol-
258
Puccinelli and Beebe
ogy due to small confined drops that limit the use of adherent cell types and limits nutrients availability [93, 94].
8.4 Drug Screening Applications The drug screening sector has primarily been centered in the pharmaceutical industry; however, with the development of microfluidic technologies and the presence of academic screening facilities, drug screening is being utilized more in university research labs as well. Cytotoxicity testing is a main avenue for drug screening. Many of the ufHTS cell culture platforms presented above have applications in both understanding the cellular microenvironment and toxicity testing. As previously described for cell culture techniques, the ability to create more native constructs for cell based drug screening is of increasing importance. Not only does the cell phenotype, expression profile, morphology, size, growth, proliferation and death rate change when in 3D culture verses 2D, but as does the cell’s possible reaction to a particular drug [95]. Developing better in vitro cell culture models results in less time and money spent on drugs that eventually fail during animal testing or clinical trials. The defined microenvironment and fluidic control present both temporally and spatially in microfluidics enable the generation of concentration gradients and the ability to treat only portions of a single cell with one condition verses another [96] making these systems appealing for drug screening [14]. Devices have been developed specifically for cytotoxicity testing. An array of 576 chambers (24x24) containing 8 u-shaped traps for cells (~10 cells/trap) screened three cell types against five toxins and was analyzed with live/dead stain via microscopy [97]. As previously mentioned, 3D culture or spheroid culture is a better model for cell based drug screening. A device was developed to form spheroids by trapping them in a similar ushaped trap with perfusion flow under the cells/traps to mimic a capillary type system [98]. The device arrays 7500 traps/cm2. While this device is a good model for cancer cell drug testing, all traps were exposed to the same fluidic environment, future designs would require the traps to be separated or the addition of an upstream gradient generator to allow for unique drug dose testing. Another group developed a method to trap and release beads or cells in a micro array for dynamic proteomics, diagnostics and drug discovery [99]. With this method they can treat the cells and then recover them after treatment. Microfluidics has been used as a tool to assist in current HTS assays by utilizing fluidic control to create serial dilutions that
High Throughput Screening Using Microfluidics
259
are dispensable into well plates [100]. Extensive work has also been done utilizing centrifugal forces in microfluidics to manipulate fluids, and these CD based platforms have been reviewed [101]. The drug screening sector is an attractive potential market for microfluidic because reduced reagent volumes reduce costs associated with expensive drugs. Microfluidics can further reduce costs by creating more defined microenvironments and 3D culture screening platforms leading to more predictive results.
8.5 Users and Developers of µF HTS Platforms The demand for high throughput screening is continually increasing across the globe. There are three main users within this sector, all with differing needs. Users include research labs, academic screening facilities, and industry listed from least throughput to the highest throughput. There often exists a gap between the developers of the platforms and ultimate end users. Currently in ufHTS, the literature reveals the majority of development within the research lab resulting in capabilities that are defined by their needs ranging from 10s to 100s of assays per day or week. On the other side of the spectrum industry has scaled to more than 10,000 assays per day for drug screening and has been standardized to well plates. In addition, with the ever increasing world population and recent boom in maize based ethanol fuels; stronger, more sustainable and higher yielding crops that are resistant to drought and insects has created a HTS seed genetics market at very high throughput. Academic screening facilities fit in the middle suiting the demands of the university, offering opportunities to screen at modest levels while encouraging the testing of new platforms. The scalability of throughput depends on quality as well as the number of assays. Using the scale of microfluidic systems (reduced sample sizes); assays can be performed with primary cells from a single source. This results in better controls and experimental consistency between various conditions. Microfluidics offers an avenue to peruse a broader range of applications for HTS by increasing control and manipulation of the smaller sample sizes.
260
Puccinelli and Beebe
8.5.1 Users: Research labs, academic screening facilities, and pharmaceutical
Few µfHTS systems have been commercialized. This is due to a number of factors, including specialization of the device for a specific application or the complexity of the system requires a base of expertise and/or specialized equipment for operation. As complexity increases with the addition of more valves, pumps, mixers and therefore more connections to outside equipment, set up time and troubleshooting for the device becomes more extensive reducing the likelihood of commercialization or use outside the lab that developed it. This is not to say, however, that these devices do not have utility. In fact, much of the work presented above has provided unprecedented insights into many relevant questions and will continue to do so in academia benefiting the life sciences. The scale of throughput in research labs is small compared to that of pharmaceutical companies. As cited here many devices offer parallel operations in the range of 10s – 100s per device. A few offer higher throughput in the 1000s. Reasons for more limited throughput in microfluidic research are non-standardized endpoints, data collection, and device real-estate to house fluidic manipulation tools. Biological endpoints are typically designed for larger volumes and standard formats. Using traditional endpoints in microfluidics often requires adaptation of the assay or device and requires comparisons to traditional methods to obtain meaningful results. This has led to an abundance of microfluidic cell culture platforms being verified with LIVE/DEAD stain only. Analysis for pharmaceutical screens that are carried out in standardized plate formats are readily scanned on various plate-readers with an automated data collection. Non-standard ufHTS systems are mainly analyzed through imaging techniques using microscopy with only a few using plate readers [55, 102]. Imaging type of a data collection generally requires a significantly more user interface reducing throughput. In addition, there is a divide between the HTS community and the microfluidic engineers, where microfluidics is rooted in the integration of devices to syringe pumps and pressure driven sources, the high throughput industry was established with automated liquid handlers and plate readers. Unfortunately, what is often available and familiar for the microfluidic lab is not for the biology lab and visa versa. Also, it is often not the goal of research labs to achieve the throughput seen in industry but rather to devise platforms having a broader parameter space as opposed to a single defined assay. Academic Screening Facilities offer a middle ground to connect ufHTS research with the resources of the pharmaceutical industry. Such, facilities exist at numerous universities offering
High Throughput Screening Using Microfluidics
261
important resources and knowledge to the research labs. Collaborations between microfluidics and biology will advance both fields. 8.5.2 Commercialized products in HTS
The research that has been able to bridge the gap between the bench top design and lab automation lends itself to commercialization. One of the longer standing companies in microfluidics is Fluidigm (http://www.fluidigm.com). They currently have two lines of products: Topaz for protein crystallization and BioMark for RT-PCR, genotyping and detection of mutant sequences. Both operate off of integrated fluidic circuit (IFC) technology. Topaz requires specialized hardware and software which they offer with their microfluidic chips and can perform 384 conditions per chip with 4 chips simultaneously per system. The BioMark’s IFC is flanked by wells which is readily interfacd with automated robotic liquid handlers requiring little additional hardware for sample loading. It can run 2,304 simultaneous PCR assays with 48 samples and 48 primer sets. Caliper Life Sciences (http://www.caliperls.com) carries a range of complete ufHTS systems for culture, assays, and analysis. Their LabChip 3000 for enzymatic and cell based drug screening has been adopted by 75% of the top 15 pharmaceutical companies. This system has been noted to not only save reagent costs in nL size reactions but increase confidence through the precision of the instrument [103]. They also offer the LabChip EZ Reader for kinase assays and a LabChip 90 for 1D electrophoretic separation of proteins and nucleic acids. All systems acquire reagents from 96 or 384 well plates and dispense them into the microfluidic network utilizing pressure or voltage to control fluid movement. These systems come complete with integrated optical equipment for automated detection. RainDance Technologies (http://www.raindancetechnologies.com) commercialized droplet based assays in microfluidic systems. Their systems can perform a wide range of applications including DNA sequencing Prep, PCR, genotyping, gene expression, FACS, enzyme screening, molecular evolution, microbiology, toxicology, RNAi, and with stem cells. Droplets from 0.5 pL to 100 nL suspended in an inert carrier oil can be moved at a rate up to 10,000 samples per second. Their Professional Laboratory System is a complete setup of hardware and software to perform all the liquid handling and analysis.
262
Puccinelli and Beebe
BellBrook Labs (http://www.bellbrooklabs.com) has commercialized an isolated channel network platform, the MCA (brand named Iuvo) [19]. The small microfluidic channels are fabricated in tissue culture treated polystyrene to eliminate problems associated with PDMS. Additionally, the technology requires no specialized equipment having channel inputs and outputs designed to fit current 96 or 384 well plate liquid handler formats for all fluidic manipulation including initial filling of the channels, seeding of cells, washing, and staining. Currently three Iuvo designs are offered; the 192 straight channel ‘Single Array’ for high throughput cellular assays providing reduced reagent volumes and improved washing over traditional well plates, the ‘Dual Array’ with two inputs and one output for co-culture applications, and the ‘Gradient Array’ for chemotaxis/migration studies. These devices have been tested with both 2D and 3D cell culture. Other companies are commercializing HTS microfluidic technologies as well. Celula Inc. (http://www.celula-inc.com) has developed personalized medicine and diagnostics devices. Advanced Liquid Logic (http://www.liquid-logic.com) is developing droplet based fluidics. SpinX (http://www.spinx-technologies.com) commercialized centrifugal fluid manipulation in a microfluidic platform controlled by a virtual laser valve for assays comparable to those in 384 well plates. Other microfluidic companies with potential for high throughput systems include Micronics (micronics.net), Fluxion Biosciences (fluxionbio.com), Eksigent Technologies (eksigent.com), and Cytoplex BioSciences (cytoplex.com). As more products become easily interfaced with current technologies reducing the barrier to entry it is expected that more platforms will reach commercialization ultimately increasing throughput in a diversified community.
8.6 Conclusion Overall microfluidics technologies offer significant potential advantages for high throughput screening. Microfluidic systems can create controlled contamination reduced environments by reducing regent volumes; providing laminar flow regimes for complex fluid networks; and allowing integration of electrodes, circuits, micropumps, and valves for greater fluidic control. Through various fabrication techniques, highly parallelized platforms have been applied to cell culture, perfusion assays, cell sorting, PCR, electrophoresis, immunoassays, synthesis, and drug screening. There have been limitations in taking these systems to market, most notable of which are: complexity of actuation, evaporation when working with small
High Throughput Screening Using Microfluidics
263
volumes and automation of cell based assay protocols as well as the time required for readouts [104]. The future applications of microfluidics in HTS are wide open as the density of assays on a device continues to increase along with increased functionality through serial assay integration. The key to the future of microfluidics in high throughput screening is the collaboration between the developers in the microfluidic community and the ultimate end users in biology and life sciences to develop platforms that are widely accepted
8.7 Acknowledgements The authors would like to thank the entire Microtechnology, Medicine and Biology Lab for their and support and for contributing their research to this chapter. D. J. Beebe has an ownership interest in Bellbrooks Labs, LLC which has licensed technology presented in this manuscript. This work was supported by NIH Grant no. R21-CA122672 and K25-CA104162-02.
References 1. Pereira DA, Williams JA (2007) Origin and evolution of high throughput screening. British Journal Of Pharmacology 152:53-61. 2. RITechnologies (2007) High Throughput Screening. RI Technologies, pp. 1181. 3. El-Ali J, Sorger PK, Jensen KF (2006) Cells on chips. Nature 442:403-411. 4. Kornienko O, Lacson R, Kunapuli P, Schneeweis J, Hoffman I, Smith T, Alberts M, Inglese J, Strulovici B (2004) Miniaturization of whole live cellbased GPCR assays using microdispensing and detection systems. Journal of Biomolecular Screening 9:186-195. 5. Larson B, Worzella T (2005) Perform Multiplexed Cell-Based Assays on Automated Platforms. Promega Cell Notes 12:13-16. 6. Purcell EM (1977) Life at Low Reynolds-Number. American Journal of Physics 45:3-11. 7. Brody JP, Yager P, Goldstein RE, Austin RH (1996) Biotechnology at low Reynolds numbers. Biophysical Journal 71:3430-3441. 8. Whitesides GM, Ostuni E, Takayama S, Jiang XY, Ingber DE (2001) Soft lithography in biology and biochemistry. Annual Review of Biomedical Engineering 3:335-373. 9. Walker GM, Zeringue HC, Beebe DJ (2004) Microenvironment design considerations for cellular scale studies. Lab on a Chip 4:91-97. 10. Voldman J, Gray ML, Schmidt MA (1999) Microfabrication in biology and medicine. Annual Review of Biomedical Engineering 1:401-425.
264
Puccinelli and Beebe
11. Pearce TM, Williams JC (2007) Microtechnology: Meet neurobiology. Lab on a Chip 7:30-40. 12. Kobayashi J, Mori Y, Kobayashi S (2006) Multiphase organic synthesis in microchannel reactors. Chemistry-An Asian Journal 1:22-35. 13. Bissonnette L, Bergeron MG (2006) Next revolution in the molecular theranostics of infectious diseases: microfabricated systems for personalized medicine. Expert Review Of Molecular Diagnostics 6:433-450. 14. Dittrich PS, Manz A (2006) Lab-on-a-chip: microfluidics in drug discovery. Nature Reviews Drug Discovery 5:210-218. 15. Chin CD, Linder V, Sia SK (2007) Lab-on-a-chip devices for global health: Past studies and future opportunities. Lab on a Chip 7:41-57. 16. Melin J, Quake SR (2007) Microfluidic large-scale integration: The evolution of design rules for biological automation. Annual Review Of Biophysics And Biomolecular Structure 36:213-231. 17. Easley CJ, Karlinsey JM, Bienvenue JM, Legendre LA, Roper MG, Feldman SH, Hughes MA, Hewlett EL, Merkel TJ, Ferrance JP, Landers JP (2006) A fully integrated microfluidic genetic analysis system with sample-in-answerout capability. Proceedings of the National Academy of Sciences of the United States of America 103:19272-19277. 18. Park TH, Shuler ML (2003) Integration of cell culture and microfabrication technology. Biotechnology Progress 19:243-253. 19. Meyvantsson I, Warrick JW, Hayes S, Skoien A, Beebe DJ (Submitted) Automated cell culture in high density tubeless microfluidic device arrays. Lab on a Chip. 20. Hung PJ, Lee PJ, Sabounchi P, Lin R, Lee LP (2005) Continuous perfusion microfluidic cell culture array for high-throughput cell-based assays. Biotechnology and Bioengineering 89:1-8. 21. Gomez-Sjoberg R, Leyrat AA, Pirone DM, Chen CS, Quake SR (2007) Versatile, Fully Automated, Microfluidic Cell Culture System. Anal. Chem. 79:8557-8563. 22. Chang CC, Yang RJ (2007) Electrokinetic mixing in microfluidic systems. Microfluidics And Nanofluidics 3:501-525. 23. Paguirigan A, Beebe DJ (2006) Gelatin based microfluidic devices for cell culture. Lab on a Chip 6:407-413. 24. Rosenthal A, Macdonald A, Voldman J (2007) Cell patterning chip for controlling the stem cell microenvironment. Biomaterials 28:3208-3216. 25. Paguirigan AI, Puccinelli JP, Abhyankar VV, Beebe DJ (2007) Using microfluidics to understand and control the cellular microenvironment. In: Gomez, FA (ed.), Biological Applications of Microfluidics. Wiley-Interscience. 26. Yu HM, Meyvantsson I, Shkel IA, Beebe DJ (2005) Diffusion dependent cell behavior in microenvironments. Lab on a Chip 5:1089-1095. 27. Lee PJ, Hung PJ, Rao VM, Lee LP (2006) Nanoliter scale microbioreactor array for quantitative cell biology. Biotechnology and Bioengineering 94:5-14. 28. Jeon NL, Dertinger SKW, Chiu DT, Choi IS, Stroock AD, Whitesides GM (2000) Generation of solution and surface gradients using microfluidic systems. Langmuir 16:8311-8316.
High Throughput Screening Using Microfluidics
265
29. Keenan TM, Hsu C-H, Folch A (2006) Microfluidic "jets" for generating steady-state gradients of soluble molecules on open surfaces. Applied Physics Letters 89:114103-114105. 30. Abhyankar VV, Lokuta MA, Huttenlocher A, Beebe DJ (2006) Characterization of a membrane-based gradient generator for use in cell-signaling studies. Lab on a Chip 6:389-393. 31. Keenan TM, Folch A (2008) Biomolecular gradients in cell culture systems. Lab on a Chip 8:34-57. 32. Huh D, Gu W, Kamotani Y, Grotberg JB, Takayama S (2005) Microfluidics for flow cytometric analysis of cells and particles. Physiological Measurement 26:R73-R98. 33. Beebe DJ, Mensing GA, Walker GM (2002) Physics and applications of microfluidics in biology. Annual Review of Biomedical Engineering 4:261-286. 34. Choi NW, Cabodi M, Held B, Gleghorn JP, Bonassar LJ, Stroock AD (2007) Microfluidic scaffolds for tissue engineering. Nature Materials 6:908-915. 35. Toepke MW, Beebe DJ (2006) PDMS absorption of small molecules and consequences in microfluidic applications. Lab on a Chip 6:1484-1486. 36. Mukhopadhyay R (2007) When PDMS isn't the best. Analytical Chemistry 79:3248-3253. 37. Castner DG, Ratner BD (2002) Biomedical surface science: Foundations to frontiers. Surface Science 500:28-60. 38. Munson MS, Hasenbank MS, Fu E, Yager P (2004) Suppression of nonspecific adsorption using sheath flow. Lab on a Chip 4:438-445. 39. Heo YS, Cabrera LM, Song JW, Futai N, Tung YC, Smith GD, Takayama S (2007) Characterization and resolution of evaporation-mediated osmolality shifts that constrain microfluidic cell culture in poly(dimethylsiloxane) devices. Analytical Chemistry 79:1126-1134. 40. Jeon NL, Baskaran H, Dertinger SKW, Whitesides GM, Van de Water L, Toner M (2002) Neutrophil chemotaxis in linear and complex gradients of interleukin-8 formed in a microfabricated device. Nature Biotechnology 20:826830. 41. Kim L, Toh YC, Voldman J, Yu H (2007) A practical guide to microfluidic perfusion culture of adherent mammalian cells. Lab on a Chip 7:681-694. 42. King KR, Wang SH, Irimia D, Jayaraman A, Toner M, Yarmush ML (2007) A high-throughput microfluidic real-time gene expression living cell array. Lab on a Chip 7:77-85. 43. King KR, Wang S, Jayaraman A, Yarmush ML, Toner M (2008) Microfluidic flow-encoded switching for parallel control of dynamic cellular microenvironments. Lab on a Chip 8:107-116. 44. Bissell MJ, Radisky DC, Rizki A, Weaver VM, Petersen OW (2002) The organizing principle: microenvironmental influences in the normal and malignant breast. Differentiation 70:537-546. 45. Paguirigan AL, Beebe DJ (2007) Protocol for the fabrication of enzymatically crosslinked gelatin microchannels for microfluidic cell culture. Nat. Protocols 2:1782.
266
Puccinelli and Beebe
46. Torisawa Y, Shiku H, Yasukawa T, Nishizawa M, Matsue T (2005) Multichannel 3-D cell culture device integrated on a silicon chip for anticancer drug sensitivity test. Biomaterials 26:2165-2172. 47. Torisawa YS, Takagi A, Nashimoto Y, Yasukawa T, Shiku H, Matsue T (2007) A multicellular spheroid array to realize spheroid formation, culture, and viability assay on a chip. Biomaterials 28:559-566. 48. Wu M-H, Huang S-B, Cui Z, Cui Z, Lee G-B (2008) A high throughput perfusion-based microbioreactor platform integrated with pneumatic micropumps for three-dimensional cell culture. Biomedical Microdevices 10: 309-319. 49. Haslam SZ, Woodward TL (2003) Host microenvironment in breast cancer development - Epithelial-cell-stromal-cell interactions and steroid hormone action in normal and cancerous mammary gland. Breast Cancer Research 5:208-215. 50. Shekhar MP, Werdell J, Tait L (2000) Interaction with endothelial cells is a prerequisite for branching ductal-alveolar morphogenesis and hyperplasia of preneoplastic human breast epithelial cells: regulation by estrogen. Cancer Res 60:439-49. 51. Medina D (2004) Stromal fibroblasts influence human mammary epithelial cell morphogenesis. Proceedings of the National Academy of Sciences of the United States of America 101:4723-4724. 52. Khademhosseini A, Yeh J, Eng G, Karp J, Kaji H, Borenstein J, Farokhzad OC, Langer R (2005) Cell docking inside microwells within reversibly sealed microfluidic channels for fabricating multiphenotype cell arrays. Lab on a Chip 5:1380-1386. 53. Walker GM, Beebe DJ (2002) A passive pumping method for microfluidic devices. Lab Chip 2:131-4. 54. Meyvantsson I, Beebe DJ (2005) 3rd International IEEE-EMBS Conference on Microtechnologies in Medicine and Biology. Piscataway: IEEE, Honululu, Hawaii, pp. 42-44. 55. Yu HM, Alexander CM, Beebe DJ (2007) A plate reader-compatible microchannel array for cell biology assays. Lab on a Chip 7:388-391. 56. Meyvantsson I (2006) Microfluidics Unplugged: Tools For In Vitro Cell Biology. PhD thesis, University of Wisconsin - Madison. 57. Puccinelli JP, Meyvantsson I, Beebe DJ (2006) 3D Microfluidic Separately Addressable Co-Culture. BMES Annual Fall Meeting. Chicago, IL. 58. Warrick J, Regeher K, Domenech M, Meyvantsson I, Wagner C, Alexander C, Beebe DJ (2007) High-throughput µFluidic Cellular Assays. The Proceedings of µTAS. Paris, France, Vol. 1, pp. 146-148. 59. Warrick J, Meyvantsson I, Ju JI, Beebe DJ (2007) High-throughput microfluidics: improved sample treatment and washing over standard wells. Lab on a Chip 7:316-321. 60. Berthier E, Beebe DJ (2007) Flow rate analysis of a surface tension driven passive micropump. Lab on a Chip 7:1475-1478. 61. Toepke MW, Abhyankar VV, Beebe DJ (2007) Microfluidic logic gates and timers. Lab on a Chip 7:1449-1453.
High Throughput Screening Using Microfluidics
267
62. Berthier E, Warrick J, Yu H, Beebe DJ (2008) Managing evaporation for more robust microscale assays. Part 1: Volume loss in high throughput assays. Lab on a Chip. DOI: 10.1039/b717422e. 63. Yu H (2006) Microenvironment Growth Control Of Mouse Mammary Epithelial Cells. PhD, University of Wisconsin - Madison. 64. Yu HM, Alexander CM, Beebe DJ (2007) Understanding microchannel culture: parameters involved in soluble factor signaling. Lab on a Chip 7:726730. 65. Pamme N, Wilhelm C (2006) Continuous sorting of magnetic cells via on-chip free-flow magnetophoresis. Lab on a Chip 6:974-980. 66. Studer V, Jameson R, Pellereau E, Pepin A, Chen Y (2004) A microfluidic mammalian cell sorter based on fluorescence detection. Microelectronic Engineering 73-74:852-857. 67. Fu AY, Chou HP, Spence C, Arnold FH, Quake SR (2002) An integrated microfabricated cell sorter. Analytical Chemistry 74:2451-2457. 68. Kovac JR, Voldman J (2006) Intuitive, Visual, Complex Phenotype Cell Sorting Using Opto-Flucs (Opto-Fluidic Cell Sorting). BMES Annual Fall Meeting. Chicago, Illinois, p. 700. 69. MacDonald MP, Spalding GC, Dholakia K (2003) Microfluidic sorting in an optical lattice. Nature 426:421-424. 70. Wang MM, Tu E, Raymond DE, Yang JM, Zhang HC, Hagen N, Dees B, Mercer EM, Forster AH, Kariv I, Marchand PJ, Butler WF (2005) Microfluidic sorting of mammalian cells by optical force switching. Nature Biotechnology 23:83-87. 71. Bessette PH, Hu XY, Soh HT, Daugherty PS (2007) Microfluidic library screening for mapping antibody epitopes. Analytical Chemistry 79:21742178. 72. Hu XY, Bessette PH, Qian JR, Meinhart CD, Daugherty PS, Soh HT (2005) Marker-specific sorting of rare cells using dielectrophoresis. Proceedings of the National Academy of Sciences of the United States of America 102:15757-15761. 73. Carbonaro A, Godley L, Sohn LL (2007) Cell Characterization Using ProteinFunctionalized Pores. The Proceedings of µTAS. Paris, France, Vol. 2, pp. 1173-5. 74. Marcus JS, Anderson WF, Quake SR (2006) Parallel picoliter RT-PCR assays using microfluidics. Analytical Chemistry 78:956-958. 75. Zhang CS, Xing D, Li YY (2007) Micropumps, microvalves, and micromixers within PCR microfluidic chips: Advances and trends. Biotechnology Advances 25:483-514. 76. Frey O, Bonneick S, Hierlemann A, Lichtenberg J (2007) Autonomous microfluidic multi-channel chip for real-time PCR with integrated liquid handling. Biomedical Microdevices 9:711-718. 77. Zhou XC, Cai SY, Hong AL, You QM, Yu PL, Sheng NJ, Srivannavit O, Muranjan S, Rouillard JM, Xia YM, Zhang XL, Xiang Q, Ganesh R, Zhu Q, Matejko A, Gulari E, Gao XL (2004) Microfluidic PicoArray synthesis of oli-
268
Puccinelli and Beebe
godeoxynucleotides and simultaneous assembling of multiple DNA sequences. Nucleic Acids Research 32:5409-5417. 78. Toriello NM, Liu CN, Mathies RA (2006) Multichannel reverse transcriptionpolymerase chain reaction microdevice for rapid gene expression and biomarker analysis. Analytical Chemistry 78:7997-8003. 79. Buch JS, Rosenberger F, Highsmith WE, Kimball C, DeVoe DL, Lee CS (2005) Denaturing gradient-based two-dimensional gene mutation scanning in a polymer microfluidic network. Lab on a Chip 5:392-400. 80. Simpson PC, Roach D, Woolley AT, Thorsen T, Johnston R, Sensabaugh GF, Mathies RA (1998) High-throughput genetic analysis using microfabricated 96-sample capillary array electrophoresis microplates. Proceedings of the National Academy of Sciences 95:2256-2261. 81. Emrich CA, Tian HJ, Medintz IL, Mathies RA (2002) Microfabricated 384lane capillary array electrophoresis bioanalyzer for ultrahigh-throughput genetic analysis. Analytical Chemistry 74:5076-5083. 82. Bilitewski U, Genrich M, Kadow S, Mersal G (2003) Biochemical analysis with microfluidic systems. Analytical And Bioanalytical Chemistry 377:556569. 83. Mangelings D, Heyden YV (2007) High-throughput screening and optimization approaches for chiral compounds by means of microfluidic devices. Combinatorial Chemistry & High Throughput Screening 10:317-325. 84. Dishinger JF, Kennedy RT (2007) Serial immunoassays in parallel on a microfluidic chip for monitoring hormone secretion from living cells. Analytical Chemistry 79:947-954. 85. Maerkl SJ, Quake SR (2007) A systems approach to measuring the binding energy landscapes of transcription factors. Science 315:233-237. 86. Situma C, Hashimoto M, Soper SA (2006) Merging microfluidics with microarray-based bioassays. Biomolecular Engineering 23:213-231. 87. LePioufle B, Suzuki H, Tabata KV, Noji H, Takeuchi S (2008) Lipid Bilayer Microarray for Parallel Recording of Transmembrane Ion Currents. Anal. Chem. 80:328-332. 88. Link DR, Grasland-Mongrain E, Duri A, Sarrazin F, Cheng ZD, Cristobal G, Marquez M, Weitz DA (2006) Electric control of droplets in microfluidic devices. Angewandte Chemie-International Edition 45:2556-2560. 89. Srinivasan V, Pamula VK, Fair RB (2004) An integrated digital microfluidic lab-on-a-chip for clinical diagnostics on human physiological fluids. Lab on a Chip 4:310-315. 90. Cho SK, Zhao YJ, Kim CJ (2007) Concentration and binary separation of micro particles for droplet-based digital microfluidics. Lab on a Chip 7:490-498. 91. Chang YH, Lee GB, Huang FC, Chen YY, Lin JL (2006) Integrated polymerase chain reaction chips utilizing digital microfluidics. Biomedical Microdevices 8:215-225. 92. Teh S-Y, Lin R, Hung L-H, Lee AP (2008) Droplet microfluidics. Lab on a Chip 8:198-220.
High Throughput Screening Using Microfluidics
269
93. Metze J (2003) Generation of larger numbers of separated microbial populations by cultivation in segmented-flow microdevices. Lab on a Chip 3:202207. 94. Grodrian A, Metze J, Henkel T, Martin K, Roth M, Kohler JM (2004) Segmented flow generation by chip reactors for highly parallelized cell cultivation. Biosensors & Bioelectronics 19:1421-1428. 95. Kunz-Schughart LA, Freyer JP, Hofstaedter F, Ebner R (2004) The use of 3-D cultures for high-throughput screening: The multicellular spheroid model. Journal of Biomolecular Screening 9:273-285. 96. Takayama S, Ostuni E, LeDuc P, Naruse K, Ingber DE, Whitesides GM (2003) Selective Chemical Treatment of Cellular Microdomains Using Multiple Laminar Streams. Chemistry & Biology 10:123-130. 97. Wang ZH, Kim MC, Marquez M, Thorsen T (2007) High-density microfluidic arrays for cell cytotoxicity analysis. Lab on a Chip 7:740-745. 98. Wu L, Di Carlo D, Lee L (ASAP) Microfluidic self-assembly of tumor spheroids for anticancer drug discovery. Biomedical Microdevices. 99. Tan WH, Takeuchi S (2007) A trap-and-release integrated microfluidic system for dynamic microarray applications. Proceedings of the National Academy of Sciences of the United States of America 104:1146-1151. 100. Bang H, Lim SH, Lee YK, Chung S, Chung C, Han DC, Chang JK (2004) Serial dilution microchip for cytotoxicity test. Journal of Micromechanics and Microengineering 14:1165-1170. 101. Madou M, Zoval J, Jia GY, Kido H, Kim J, Kim N (2006) Lab on a CD. Annual Review of Biomedical Engineering 8:601-628. 102. Weigl BH, Morris CJ, Kesler N, Saltsman P, Bardell R (2001) Well-plate formats and microfluidics–applications of laminar diffusion interfaces to HTP screening. In Ramsey, JM and A. van den Berg, (eds.), Micro Total Analysis Systems. Monterey, CA, Vol. 1, p. 383–384. 103. Boguslavsky J (2004) Is Microfluidics Equipped for HTS. Bio-IT World. 104. Comley J (2005) New Options for Cell-Based Assay Automation. Drug Discovery World Fall:39-62.
Chapter 9 Microfluidic Diagnostic Systems for the Rapid Detection and Quantification of Pathogens
Shramik Sengupta1, Jason E. Gordon2, and Hsueh-Chia Chang3 Department of Biological Engineering; University of Missouri, Columbia, MO 65211 1
2
Special Programs Division; Midwest Research Institute, Kansas City, MO 64110 Center for Microfluidics & Medical Diagnostics; Department of Chemical & Biomolecular Engineering; University of Notre Dame, Notre Dame, Indiana 46556
3
Correspondence should be addressed to: Hsueh-Chia Chang (
[email protected])
Keywords: bacteria detection, virus detection, microbiome, rapid diagnostics, point-of-care diagnostics, biomedical diagnostics, microfluidics
Abstract This article reviews past and current research directed towards developing microfluidic systems that are able to rapidly detect the presence of pathogens and provide additional clinically relevant information about them (e.g., their antibiotic susceptibility, etc.) about them. It is estimated that
272
Sengupta, Gordon, and Chang
pathogens are directly responsible for 15 million deaths worldwide annually. Many of these deaths could be prevented as a result of rapid and/or point-of-care diagnosis. The microfluidic systems reviewed here , which are in use and under development, seek to fulfill this significant need. The technical objective of these systems is to detect and estimate the concentration of pathogens of interest within clinical samples. This can be achieved by the detection and quantification of any of the following: (a) whole pathogen cells; (b) metabolites released or consumed by the pathogen; and (c) proteins or nucleic acid sequences that are specific to the pathogen of interest. The type of target assayed forms the basis behind the classification of the wide varieties of microfluidic systems encountered. However, irrespective of the actual target assayed for, the microfluidic systems have to overcome twin problems: that of low pathogen concentration and/or the presence of interferents. Hence various strategies, such as culturing, filtration, electro-kinetic separation, magnetic separation, enzyme linked immunosorbent assay (ELISA), polymerase chain reaction (PCR), etc., are used to amplify and/or isolate the target pathogen prior to detection. Detection is accomplished using optical, electrical and mechanical techniques (or a combination of them). In this review, attempts have been made to highlight how the physics at the micro-scale has influenced the design of the separation and detection schemes employed.
9.1 Introduction 9.1.1 Infectious pathogens and their prevalence Infectious pathogens have posed a considerable threat to global health throughout history. Such pathogens are responsible for the black death (caused by Yersinia pestis) of the fourteenth century and the influenza pandemic that occurred between 1918 and 1920, each of which claimed more than 50 million lives [1-3]. More recently, the acquired immunodeficiency syndrome, AIDS (caused by the human immunodeficiency virus, HIV), was first recognized in 1981, and severe acute respiratory syndrome (SARS) emerged in 2003. Additional examples of infectious diseases and their causative agents include hemolytic uremic syndrome and bloody diarrhea (Escherichia coli O157:H7), tuberculosis (Mycobacterium tuberculosis), anthrax (Bacillus anthracis), pneumonia (Streptococcus pneumoniae and Avian influenza virus), malaria (Plasmodium), hepatitis (Hepatitis A,
Microfluidic Diagnostic Systems
273
B, C, D, and E virus), and hemorrhagic fever (Ebola virus) amongst many others (see Figure 9.1). More than 50 new, or newly discovered, pathogens have been identified over the past 30 years [1]. In addition to these new discoveries, there exists the constant threat of re-emerging pathogens, which often occur throughout new or extended geographic regions. West Nile fever is one infectious disease that has recently (1999) made a reappearance within a new geographical location, the United States [4]. This disease is the result of a flavivirus transmitted via mosquitoes. Bioterrorism and the growing phenomenon of antimicrobial resistance further compound the difficulties posed by infectious pathogens. One familiar example of a bioterrorist attack took place in 2001 with the mailing of letters filled with anthrax spores that ultimately claimed the lives of 5 people and caused illness in 13 others [6]. The issue of antibiotic resistance is not necessarily new, but, is growing for a variety of reasons, including the inappropriate utilization of antimicrobials [7, 8]. Infectious pathogens that have developed antimicrobial resistant strains over the last several years include pneumococci, enterococci, staphylococci, Plasmodium falciparum, and Mycobacterium tuberculosis [5].
Fig. 9.1 The global prevalence of infectious diseases. Reprinted with permission from Macmillan Publishers Ltd: Nature [2], copyright 2004; and from the University of Chicago Press: Clinical Infectious Diseases [5], copyright 2001.
274
Sengupta, Gordon, and Chang
As can be seen from Figure 9.1, the prevalence of infectious diseases is a global phenomenon. It is estimated that approximately 25% (15 million of 57 million) of annual worldwide deaths are directly attributable to infectious diseases [2]. The toll of these diseases is much greater in the developing world than it is within the developed world [9], with at least half of all deaths being caused by infectious pathogens [10]. However, even within developed countries, microbial pathogens are not a trivial public health issue. An appreciation of their significance can be gained by examining relevant sepsis statistics. Sepsis is a condition that occurs when toxinproducing bacteria, and other pathogens, infect the bloodstream. These pathogenic organisms are often introduced into the blood stream via infected organs, such as the kidneys (upper urinary tract infection) and the lungs (bacterial pneumonia), or during a hospital stay via various routes (e.g. intravenous lines, surgical wounds, surgical drains, and open-skin wounds like bedsores). Although it does not attract much publicity, sepsis is one of the leading causes of death in the US, with an estimated 750,000 people annually developing severe sepsis and over 215,000 (close to 30%) of those people succumbing to the infection [11]. Within the US, sepsis claims more lives annually than myocardial infraction [12], lung cancer, or breast cancer [13]. Perhaps a greater cause of concern is that, unlike most other leading causes of death, the fatality rate associated with sepsis has not decreased significantly over the last few decades [14]. Prevention and early detection leading to prompt treatment are essential since minor infections can rapidly turn life threatening [15]. Multiple studies have shown that the earlier bacteria can be detected within the bloodstream and the appropriate antibiotic can be delivered, the more successful the treatment is in fighting off the infection [16-18]. Diagnostic systems therefore play an integral role in facilitating an effective response to be provided against these infections.
9.1.2 Traditional pathogen detection methods The ability to detect infectious pathogens is vital to mounting a quick and effective response to outbreaks of pathogen-caused disease. The clinical microbiologist, nurse, or technician is thus often tasked with answering the question, “Are there pathogens present in my sample (be it a clinical sample, food sample, environmental sample, etc), and if so, what is/are their identity/identities, and in what number are they present?” In general, the answer to the first part (what identities) may vary from just a few (< 5) to hundreds, and the number of each type may independently vary over sev-
Microfluidic Diagnostic Systems
275
eral orders of magnitude (from a few to millions of colony forming units (CFU) per ml of sample). The traditional, and most general, technique designed to answer this catchall question, is the plate-culture method [19]. This method was first popularized by Robert Koch in the last quarter of the 19th Century, and is still extensively used today with only minor modifications. The standard procedure is to dilute the available sample serially over a few orders of magnitude and to spread approximately 0.1 ml of each of these serially diluted sub-samples uniformly over the surface of a petridish (usually 3-4 inches in diameter) containing a bed of agar gel. This gel contains nutrients that support the growth of microorganisms. Once the liquid suspension is spread out, bacteria in the (sub)sample adhere to the agar gel. The plates are then kept at a temperature favoring growth (typically 37ºC), and after an incubation period during which bacteria on the agar proliferate, bacterial colonies that can be visually observed are formed (assuming the initial presence of target bacteria). The number of colonies established is then counted. For a limited range of dilutions of the original sample, it is possible to distinguish individual colonies, and these counts are used to establish the concentration of bacteria in the original sample. The plate-culture technique is relatively simple and provides a snapshot of the number of bacteria in the original sample at the time that it was collected. However, it does have a number of drawbacks. To begin with, the plating technique utilizes a significant amount of material, is tedious, labor intensive, and takes a long time to provide results (depending on the species and the medium, the time required may range from overnight, to days, or to weeks). The process becomes even more labor intensive if one seeks to screen a sample for multiple bacteria. In this case, multiple petridishes are required and selective media (a medium that supports the growth of only a limited class of bacteria, sometimes even just a single species) have to be designed for each of the target pathogens. Given that in many cases it is not possible to design growth media that are selective for all the individual targets, differences in colony morphology have to be relied upon in order to distinguish different targets and obtain an estimate of their concentrations in the sample. Unfortunately, identifying bacteria based on colony morphology requires very specialized training, and the readings obtained are subject to errors in observer judgment. A more subtle drawback is that the blend of nutrients present in the agar gel may not support the growth and colony formation of all the different types of bacteria present in the (sub)sample spread on the plate. It is also possible that different types of bacteria present may grow/proliferate at very different rates on the agar. It
276
Sengupta, Gordon, and Chang
is likely that non-growing or slow-growing bacteria may not be seen, in which case they will not be considered. For many applications, this is a significant limitation. For instance, it is believed that a large fraction of microbes residing in the intestines of humans and other mammals, whose absolute and relative numbers have clinical significance, have never been cultured on plates [20]. An analogous technique used to determine the viral count of samples is to first grow a “lawn” of bacteria/eukaryotic cells on a petridish, and then spread onto it a suitably serially diluted sample suspected to contain virus particles. As the virus infects and subsequently lyses the cells, clear zones or “plaques” are created in the lawn. Each plaque is analogous to individual colonies of bacteria. However, as is often witnessed with bacterial culture, not only do different viruses vary in their ability to form plaques in the cells used to make lawns, but even different strains of the same virus may also exhibit such differences [21]. The inherent limitations of culture-based detection methodologies have resulted in interest directed at, and effort devoted to, developing alternate pathogen detection and quantification methods. The main goals of these developed platforms are to speed-up the detection process and to improve accuracy. These aims have resulted in differing technologies that employ tools such as flow cytometry, standard micro-well plate enzyme-linked immunosorbent assay (ELISA), and nucleic acid based techniques. Most diagnostic devices to date, with a few exceptions, have been designed for use within the developed world with well-equipped laboratories and trained technicians [10]. However, there is now a consumer-driven demand for rapid and point-of-care (POC) diagnostic devices, which culture-based and other commonly employed detection techniques will not be able to meet due to long detection times and a lack of portability. Considering the scarce resources that characterize the developing world, such technologies also will not suffice in these regions. In these regions, a detection device must be portable, robust, rapid, and simple to operate.
9.1.3 Microfluidic techniques Microfluidic technologies are ideally suited to meet the challenges described for the rapid and POC detection of pathogens. While the goal of portability is perhaps the leading cause behind the development of microfluidic devices, such systems offer many other advantages. These include reduced detection times, reduced biological sample (and reagent) require-
Microfluidic Diagnostic Systems
277
ments, integrated devices that combine sample processing with detection on a single chip, and channel geometries and dimensions that often aid in the isolation and/or detection of the target pathogens. There exist a large number of microfluidic diagnostic systems that employ different approaches to rapidly establish the presence of the pathogen of interest in a sample of constrained size. In many cases, these techniques also estimate target pathogen concentrations to varying degrees of certainty. In general, these techniques achieve their objective by detecting/measuring any of the following: (a) whole pathogen(s) (b) metabolites released or consumed (c) genetic material or proteins specific to the target pathogen These differing types of targets provide one way of classifying the large number of microfluidic systems either in the market or under development. The next section of this review, which reports on developments in the recent past, has been divided accordingly.
9.2 Review of Research 9.2.1. Pathogen detection/quantification techniques based on detecting whole cells In “real world” samples, the concentration of whole bacterial cells is often low, and/or there are a large number of other interfering species present, complicating pathogen detection and quantification. For instance, at the time when patients begin to manifest clear clinical symptoms of sepsis, they often have bacterial loads of only 1-30 CFU (Colony Forming Units)/ml, and even for patients who are near death, the load may still be as low as ~1000 CFU/ml [22]. In addition, these low numbers of bacteria need to be detected in the presence of blood cells, of which there are millions per milliliter. Microfluidic diagnostic systems thus have to deal with two distinct, but inter-related problems, increasing the concentration of the target pathogen cells (in absolute terms and/or with respect to that of the other cells) and detecting/quantifying their presence. The easiest way to approach this overall problem is to tackle these two challenges individually. First, isolate the target cells and second, devise a platform for detection and quantification. It is expected that during this isolation process, not
278
Sengupta, Gordon, and Chang
only will the absolute concentration of targets increase, but that interfering species will be discarded as well. In many cases, especially those in which isolation of the target species from the sample medium is desired, the specificity of antibody-antigen interactions can be leveraged to enhance the separability of the target species from the sample medium. Antibodies are proteins synthesized by the immune system of animals (including humans) as a targeted response to an infection by a foreign body or pathogen (typically referred to as an antigen). The conformation of these proteins is such that they bind to specific proteins on the surfaces of the invading pathogens. Since different strains of pathogens often exhibit different proteins on their surfaces, it is possible to generate antibodies that are strain specific -- such as those developed for Escherichia coli (E. coli) O157:H7, a particularly virulent strain of E. coli [23]. It is also of importance that only a small part of the antibody molecule actually binds to the target antigen. The other end of the molecule, the terminal end, can be chemically modified and immobilized on to a wide variety of surfaces like glass or silica (silicon wafers with a thin oxide layer or glass) [24], polymers such as acrylates [25], phthalates [26], polyethylene [27], and metals such as gold [28]. As will be discussed later, once target pathogens bind to their corresponding immobilized antibodies, various fluidic, electrokinetic, and other schemes may be used to both quickly isolate them and to rapidly estimate their number. 9.2.1.1. Isolating pathogens
The following methods (culturing, filtration, magnetic, and electrokinetic) are some of the more common techniques utilized to isolate samples prior to attempts at detecting pathogens present within a given sample. Culturing techniques If the target pathogen is a viable organism, an increase in its number can be facilitated via reproduction by supplying a balanced set of nutrients. Traditional microbiology research has provided standard growth media formulations for wide-ranging bacteria cultures, making it relatively simple to reproduce a bacterial pathogen. There are also special media formulations that selectively support the growth of particular classes of microbes [29]. One such example is Eosin-Methylene Blue (EMB) agar, which is toxic to gram-positive bacteria allowing the growth of only gram-negative bacteria.
Microfluidic Diagnostic Systems
279
Most bacterial cells are slightly heavier than water and other commonly used aqueous growth media, and so the cells will sediment. During aqueous culture, standard practice is then to keep the bacterial cells suspended in the growth medium using shakers, stirrers or other forms of agitation [30]. This also serves to facilitate the transport of oxygen and nutrients to the surface of the cell, and prevents toxic wastes excreted by the cells from building up near their immediate vicinity. Additionally, the cultures are usually maintained at a constant temperature (typically 37ºC) by housing the growth vessels within incubators. Achieving similar local flow regimes and temperature profiles around proliferating cells maintained within reactors at volumes on the order of micro-liters is a non-trivial task. To resolve these challenges, microfluidic valves, mixers, and heaters have to be incorporated to culture the cells [31-34]. Additionally, sensors to measure and regulate temperature, dissolved oxygen concentration, and pH within the micro-reactor have to be developed and integrated. One such microreactor is shown in Figure 9.2.
Fig. 9.2 An example of a microfluidic system used for culturing cells. Besides the micro-reactor used to maintain the cells, the system also contains micro pumps, micro check-valves, reservoirs (for growth medium and buffers), feed channels, a micro temperature sensor, and micro heaters. Reprinted with permission from the Institute of Physics Publishing Ltd.: Journal of Micromechanics and Microengineering, [32]. Copyright 2007.
Additional challenges in micro-reactor development include considerations relating to the volume, scales, and fabrication materials typically encountered in microfluidic systems. For instance, polydimethylsiloxane (PDMS), a material commonly employed in biomicrofluidic systems, is permeable
280
Sengupta, Gordon, and Chang
to air. This material property may lead to sample losses due to evaporation [35]. While the rate may be low (micro-liters per hour or lower), the sample losses cannot be neglected when the operation volume is on the order of micro-liters and the reactor is to be maintained for hours. Potential corrective measures available include housing the microfluidic device in a humidity controlled box [36], or specially reformulating the material used to reduce losses due to evaporation [37]. Another issue of concern involves the use of a single micro-reactor for culturing multiple samples, which may be required depending on the cost of the microreactor. Under these circumstances, the formation of biofilms can be problematic. Biofilms are formed when bacteria and other microorganisms adhere to solid surfaces in contact with water and secrete a complex polymeric substance that both anchors them to the solid surface and protects them from bacteriocidal substances in the water [38]. Such biofilms will intermittently shed cells housed within them, which can lead to contamination of the samples being cultured. Biofilms are also difficult to get rid of due to the nature of the polymer matrix in which the microbial cells are housed. The challenges associated with biofilm formation are much more acute in microfluidic systems due to the high surface area to volume ratios encountered [39], and consequently using disposable systems is preferred. If for some reason a disposable system cannot be employed, complex bioreactor designs may be required to avoid biofilm formation. One such design involves the continuous transport of small, isolated volumes within a micro-channel loop [40]. As described above, the replication of macro-scale bacterial growth conditions within microfluidic devices is a non-trivial task. Therefore, in designing a diagnostic system that requires pathogen growth through culture, the potential benefits of micro-reactor culture should be weighed against the inherent challenges. Filtration techniques Filtration is conceptually a simple method for increasing the concentration of target species. The sample suspension being investigated is allowed to pass through a filter whose pore size is smaller than that of the target, resulting in the target pathogen being retained in the excluded filter cake (retentate). The retentate may then be resuspended in a volume of liquid much smaller (1/10th or even less) than the original sample. At the macroscale, filtration is more commonly employed for applications like environmental monitoring and food safety, where large sample volumes (1L or larger) and low bacterial loads (~1 CFU / 100 ml or lower) are commonly
Microfluidic Diagnostic Systems
281
encountered. On the other hand, clinical samples are typically much smaller. Given the limited aliquot size that can be introduced into a microfluidic device, the use of filtration at the micro-scale may also be necessitated. For example, consider the case of a liquid clinical sample with a bacterial concentration of 100 CFU/ml. If the aliquot size to be employed within the microfluidic device is 100 µl, then it can be expected that the aliquot will contain 10 CFU on average. However, if the aliquot size is further reduced to 10 µl, then the average colony count within an aliquot will only be 1 CFU, meaning there is a significant chance that the aliquot drawn will contain no bacteria (even though it was present within the original sample suspension). The ability to concentrate the bacteria from the original sample suspension through filtration would thus be beneficial in preventing false negative test results. Fabricating filters within microfluidic systems poses several challenges, which have been discussed elsewhere [41]. In brief, three types of filters are typically found; arrays of pillars within a micro-channel that are usually arranged in a regular geometric pattern [42, 43], micron-sized “holes” along one of the walls (usually the “floor”) of the channel [44, 45], and a micro-channel embedded porous matrix that typically provides a distribution of pore sizes [46-49]. A single filter with a modal pore size smaller than that of the target pathogen will retain the target, which may then be subsequently assayed [50]. However, if there are too many other non-target species larger than the pore size, this may lead to the system quickly fouling, which complicates subsequent target detection within the filter cake. This issue can be resolved through the on-chip use of two filters, one with a pore size slightly larger than the target pathogen and the other with a pore size slightly smaller than the target pathogen [51]. In this case, the material trapped between the two filters is likely to contain a higher concentration of target particles. It is still possible that a rapid buildup of filter cake on the upstream filter could occur, however, the use of cross-flow and electrokinetic stirring [52] can be employed to remove the cake. Magnetic techniques The use of magnetic techniques to isolate biological samples is a familiar technique on the macro scale [53, 54], and is potentially attractive for use within microfluidic systems for the isolation and manipulation of target pathogens. Of particular interest is the magnetic susceptibility of a particular cell [55], or of a magnetic particle that can be attached to a particular cell [56-58]. An example of a cell that exhibits magnetic susceptibility is
282
Sengupta, Gordon, and Chang
the red blood cell. Deoxygenated red blood cells are the most paramagnetic cell within the body as a result of the free electrons associated with the iron atom contained within the heme ring of hemoglobin. A continuous separation of red blood cells from whole blood has been achieved in a microfluidic system by placing a small ferromagnetic wire along the length of a microchannel and applying a uniform external magnetic field to create a high gradient magnetic field locally [55]. In the majority of cases of interest, the target pathogen will not be naturally magnetic. Consequently magnetic particles with the capability of selectively binding to the target pathogen must be employed, and there exist many examples of their use within the literature [56-59]. Such particles are advantageous for a variety of reasons, including that they offer a considerable surface area for pathogen capture, and that magnetic handling is generally not impacted by changes in pH, temperature, etc. [60]. Magnetic particles are typically composed of an interior magnetic core and an outside shell that can be made of substances ranging from polystyrene to silica to polysaccharide. The idea fundamental to the use of magnetic particles is that some form of interaction between the particle and the pathogen can be created. One prominent example relies upon the conjugation of antibodies specific to a given pathogen onto the surface of the magnetic particle. After the introduction of a sample containing the target pathogen, this pathogen will bind to the conjugated capture antibodies while other substances within the sample should remain unattached. This specific technique was employed by Lien et al in an effort to isolate Dengue virus serotype 2 and enterovirus 7 [58]. This isolation employed Dynabeads® (Dynal Biotech, Norway), which are hydrophobic superparamagnetic polystyrene beads of approximately 4.5 µm in diameter. After allowing for the target virus pathogen to bind to its respective antibody on the surface of the beads, the beads were then isolated via a magnetic field created using an integrated magnet of microcoils. While holding the beads in place, the remaining suspension could be flushed from the region, allowing for the isolation and concentration of target pathogen. As described above, the magnetic isolation of a cell or magnetic particle is reliant upon the induction of a magnetic field within the vicinity of the cell or magnetic particle to be trapped. The selection of an appropriate magnet to fulfill this challenge is thus of importance, and in general, either permanent magnets or electromagnets are employed. To date, most work has relied upon supplying a magnetic field from outside the microchannel or microchamber [60]. The development of integrated miniature permanent magnets and electromagnets has progressed and many examples of their
Microfluidic Diagnostic Systems
283
use can be found in the literature [59, 61-63]. However, improving the performance of these internal magnets remains an ongoing endeavor. For a detailed discussion of magnetic theory and magnet integration with microfluidic chips, the reader is directed towards several comprehensive reviews that also provide excellent discussions on the implementation of magnetic nano- and microparticles within lab-on-a-chip designs [60, 64, 65]. Electrokinetic techniques Electrophoresis and dielectrophoresis are the two main available electrokinetic techniques for pathogen isolation. These technologies rely on both the surface charge and the dielectric properties of bacterial cell membranes to isolate the pathogen target from the matrix or suspension in which they are dispersed. Electrophoresis refers to the migration of charged particles suspended in a fluid down the potential gradient of an externally applied electric field. Typically, particles, which are broadly defined here to include both biological cells and macromolecules such as proteins and nucleic acids, migrate with a constant velocity that arises when the viscous drag forces on them equal that of the coulombic forces propelling them down their potential gradients. The velocity of the particle per unit field gradient is referred to as its “electrophoretic mobility”. The net charge that the particle carries can be a function of the pH of the solution (most proteins and many cells have a net positive charge at low pH and a net negative charge at high pH), and the pH value at which the particle carries no net charge is referred to as its isoelectric point (pI). Particles with different electrophoretic properties can be separated from each other using various schemes, yielding different techniques such as “field flow fractionation”, “zone electrophoresis”, “isoelectric focusing”, etc. that have been described elsewhere [66, 67]. It has also been known for some time that when suspended in common buffers, bacteria of different species have different electrophoretic mobilities, and that their isoelectric points are also different [68]. These differences have been attributed to variations in the surface composition (surface charge density) between the different species/strains of bacteria [69]. Such differences in the electrophoretic behavior of cells have been extensively used at the macro-scale to separate various types of biological cells from each other and from non-biological material [70]. Capillary electrophoresis (CE) is a form of electrophoresis where the separands migrate, along with the buffer in which they are suspended, from one end of a capillary to the other due to electro-osmotic flow produced by an applied direct current (DC) voltage. Due to the electrophoretic velocity of the seperands relative to the buffer, the seperands elute at a dif-
284
Sengupta, Gordon, and Chang
ferent time than when they would have had they flowed along with the buffer. By definition, CE is a “microfluidic” technique. Though traditionally performed within specially constructed capillaries, it can also be performed within channels housed in microfluidic chips [71]. The narrow dimension of the capillary (or micro-channel) plays a key role in the success of the technique by preventing the development of recirculation vortices that tend to arise due to temperature differences generated by the joule heating of the ionic current through the system. The Grashof Number, which gives the ratio of buoyancy forces to viscous forces on a fluid element, predicts the formation of natural convection vortices, and scales as the cube of the characteristic length. Consequently, reducing the characteristic length of the system suppresses natural convection that is driven by temperature induced density differences. The high surface area to volume ratio within a microfluidic device also enables relatively rapid heat transfer between the capillary interior and the ambient, thereby further preventing the onset of natural convection. CE has been used extensively to sort “bioparticles”, such as viruses, bacteria, eukaryotic cells, and even sub-cellular organelles. Detailed reviews are available that summarize the results obtained by various researchers in this area [72, 73]. Due to their varied electrophoretic mobilities, different types of bacterial cells elute at different times from the capillary. The species being eluted can be detected using methods such as UV absorbance or laser induced fluorescence [74]. Capillary zone electrophoresis (CZE), which is the most commonly employed form of CE, is characterized by the maintenance of a constant pH along the whole separation pathway. It has been used to separate significantly differing bioparticle types, such as gram positive and gram negative bacteria [75, 76]. A more sophisticated form of CE, known as capillary isoelectric focusing (CIEF), establishes a pH gradient along the separation path and has been employed in the more difficult separation of bacteria of similar sizes [77]. There are drawbacks, such as lowered throughput, to the use of CE. Another significant disadvantage with CE stems from the fact that extremely high voltages (kilo-Volts) must be used to drive the electro-osmotic flow (EOF) through the capillaries. This high voltage requirement not only necessitates the use of bulky (and dangerous) high voltage supply units, but also produces extensive electrolysis (gas bubbles). Special provisions must then be included within the device to allow for the escape of air [78], while also compensating for the fact that if air bubbles do enter a micro-channel or capillary, then EOF alone may not be sufficient for displacing the bubbles within the conduit [79].
Microfluidic Diagnostic Systems
285
To avoid the drawbacks inherent in using DC methods such as electrophoresis, a method like dielectrophoresis (DEP), that requires the use of alternating current (AC) rather than DC voltages can be employed. In contrast to DC, within AC the polarity of the electrodes changes once every halfcycle of the AC signal. If frequencies higher than ~10 kHz are used, the half-cycle time period is shorter than the charge transfer time at the electrode for most electro-chemical reactions, and consequently most Faradaic reactions, which produce air bubbles, are suppressed [80, 81]. In characterizing DEP, it is convenient to consider it the AC analog of electrophoresis. DEP is a process in which migration of suspended particles (not necessarily charged) occurs in the presence of an electric field gradient, either towards or away from regions of high electric field strength. The migration of a particle towards a high-field region is termed positive DEP, while the opposite is known as negative DEP. Whether the particle experiences positive or negative DEP is determined by the sign of the Claussius-Mossotti factor, a complex term that depends on the values of the conductivity and dielectric constants (permittivity) of the particle and the suspending medium [82]. At low frequencies (ω<1 KHz), a particle’s dielectrophoretic behavior is almost completely dominated by differences in conductivity, while at high frequencies (ω>100 MHz) its DEP behavior is dictated by differences in permittivity. However, for frequencies that can be practically realized, ~10 kHz to 100 MHz (high enough not to generate bubbles, but low enough to not require expensive equipment), DEP behavior depends in a complex manner on both of these quantities. Qualitatively though, the effect of conductivity is more pronounced at lower frequencies and that of permittivity at higher frequencies. In standard buffers, such as phosphate buffered saline (PBS), both the conductivity and the permittivity of bacterial cells are typically lower than that of the medium. Consequently, bacterial cells generally do not display any cross-over from negative to positive DEP (have negative DEP for all frequencies). However, when these cells are placed in low conductivity suspensions (~100 mS/cm), they display positive DEP at lower frequencies and cross-over to negative DEP in the range of 100 KHz to 100 MHz [83]. Since dielectrophoretic phenomena rely on field-induced dielectric polarization at the cell membrane, and the composition and properties of the membrane vary from species to species, the cross-over frequency (the particular frequency at which the DEP behavior changes from positive to negative) is generally a function of the species and strain of the cell [84]. Thus, there can exist certain frequencies where the DEP behavior of target cells is different from that of other cells (and non cellu-
286
Sengupta, Gordon, and Chang
lar impurities in the system), allowing for the isolation of the target cells from complex suspensions containing many other species [85-88]. However, it must be considered that prolonged storage, or other sample handling processes that impact the physical composition of the membrane can change the cross-over frequency of the target cell [89, 90]. Perhaps the main disadvantage associated with DEP is that typical obtainable particle velocities are low. It can be shown that the DEP force on a particle scales to the third power of the particle radius and quadratically with the applied voltage. This cubic dependence on particle radius in conjunction with practical limitations on the applied voltage render the particle velocity produced by a DEP force miniscule, typically on the order of 10 µm/sec for bacteria and 1 µm/sec for viruses. Concentration times on the orders of hours are thus required. Also problematic is the fact that the field gradient necessary to drive DEP motion can only be achieved with relatively narrow inter-digitated electrodes, whose field penetration depth is limited by the electrode width. As a result, DEP channels are usually less than 50 microns in transverse dimension. The slow capture and the small transverse dimension produce extremely low throughputs for continuous flow kits. For example, for a 5 cm-long device employing a 50 micron transverse dimension and a 10 micron/s DEP velocity, pathogen capture can only be ensured if the throughput is less than 25 nanoliters per second. At this flowrate, it would require more than 10 hours to process a 1 ml sample. This low throughput and batch volume can be somewhat alleviated with a massively parallel array for laboratory use, but such arrays cannot be easily designed for a miniature and portable diagnostic kit. Bacteria isolation and concentration by DEP can be sped-up by using fluidic forces (convection) to transport suspended particles (of all types, including the target pathogen) to a specific localized region, and then capturing the targets utilizing DEP forces. This basic idea has been extensively applied in a variety of manners to capture different targets by the corresponding author of this review and collaborators. This can be accomplished by generating converging flows with stagnation points, and imposing the electric field gradient on the particles at the electrode stagnation points (or lines) where the viscous drag is weakest. In such systems, while all suspended particles are swept towards the stagnation point along with the fluid flow, only particles (such as the target pathogen) that are directed by the electric field (due to positive or negative DEP) towards the stagnation points are captured. Converging flows of this nature have been generated by AC EOF at symmetric electrodes [91], by AC EOF generated with serpentine coils [92], by rotational vortices generated using an ionic wind
Microfluidic Diagnostic Systems
287
[93], and by a combination of DC bias and circulatory AC EOF [94]. It is possible to generate fluidic velocities of approximately 10 mm/s, which represents a thousand-fold increase in the speed at which target particles move towards the collection zone compared to simple DEP. Once collected at a “point” (a volume of a few pico-liters), the pathogens can then be detected via a variety of detection mechanisms.
Fig. 9.3 An integrated chip that uses 3-D electrodes to sort pathogens into 3 different classes on the basis of their individual DEP cross-over frequencies. A zoomed in image of the four stages on the integrated chip is provided in (a), while (b) shows that the stages are housed within two glass slides which are detached in this image for clarification. Finally, a schematic for the trapping electrodes that each trap a different respective target particle by their negative DEP mobilities is given in (c). Reprinted with permission from Biomicrofluidics [95], copyright 2007, American Institute of Physics.
The rapid dielectrophoretic isolation and/or concentration of pathogens within continuous flow-through systems has been accomplished utilizing methods such as integrated “3-dimensional” (3D) electrodes. Such designs employ aligned electrode pairs fabricated either along the sidewalls [96] or on the top and bottom [97, 98] of channels through which the particleladen fluid flows. This 3D configuration produces a high field normal to the flow direction and penetrates across the entire height of the channel. Therefore, bacteria present through the entire depth of the channel (as opposed to just those near the bottom, as is the case for single electrode devices) can be deflected and/or captured while liquid flows unimpeded. Through tailoring the orientation of the 3D electrode, the particles can be directed to different streamlines/exits, or can be prevented from flowing past a point in the channel altogether, creating a DEP force field cage through which particles can not pass [99-102]. DEP fields from 3D electrodes have been exploited to generate a continuous sorting and trapping
288
Sengupta, Gordon, and Chang
device, which is shown in Figure 9.3, that is capable of sorting and collecting three different types of pathogens at a rate of ~300 particles/sec [95].
Fig. 9.4 Actual and fluorescent images of fluorescent nano-particles trapped between two electrodes at 1 MHz. (a) and (b) are without CNTs, (c) and (d) are with pre-assembled CNTs, and (e) and (f) are with a dispersed suspension of particles and CNTs. Reproduced with permission from [103].
The unique physical and electrical properties of nano-particles, such as carbon nano-tubes (CNTs), can also be harnessed to isolate target species dispersed in the sample. Due to their ability to store charge and their long, narrow geometry, very strong dipoles are induced by exposing CNTs to a high frequency AC field. This focused field, especially at the tips of the CNTs, leads to high field gradients, which serve to attract pathogen particles to the CNTs by DEP. CNTs may be embedded onto the “floor” of the micro-channel [104], or may be deliberately dispersed into the pathogen containing suspension [103]. In the latter case, on average the CNTs are closer to the target pathogens. Additionally, at lower AC frequencies, CNTs are themselves attracted to electrodes embedded within the microfluidic device. Their cylindrical geometry confers upon them a much higher DEP mobility than is typically observed for spherical or ellipsoidal pathogens. Consequently, suspension dispersed CNTs will drag along the
Microfluidic Diagnostic Systems
289
surface of embedded electrodes as they flow through a micro-channel, thereby bringing trapped pathogens to the electrode surface. This approach, as illustrated in Figure 9.4, has been successfully utilized to enhance the trapping rate of target pathogens by a factor estimated to be over 1000. 9.2.1.2 Estimating pathogen load
Electrical techniques In an aqueous solution, charge is carried between any two electrodes by ions in the solution. The presence of particulate matter (such as pathogens) physically obstructs the movement of these charge carrying ions, and thereby leads to a higher resistance (or impedance, when using an AC signal) between the electrodes. Typical pathogen loads (~1000 particles/ml) have a volume fraction of ~10-12, and consequently this change in resistance/impedance is not significant. It may, however, be discernable if the suspension is made to pass through a narrow slit, only slightly larger than the pathogens, that possesses electrodes on two opposite ends [105], or if the target pathogens can be made to adhere to or congregate at a surface on or very near an electrode. Quite often, the surface of interest may be treated to enhance the adhesion of cells [106], or specific antibodies may be immobilized to capture (and hence detect) only the target pathogenic entities [107, 108]. The above mentioned electrical detection techniques suffer from drawbacks that make it difficult for them to handle “real world” samples. The former, passing the suspension through a narrow slit with electrodes on two opposite ends, is plagued by the high likelihood of the slit being clogged by random particulate matters present in field samples. The latter method, forcing the target pathogens to adhere or congregate at a surface on or near an electrode, suffers from two main drawbacks. First, particles other than the target may get deposited or adhere non-specifically to the surfaces being monitored, thereby yielding misleading results. Second, the targets may take a long time to diffuse to the surface, thereby degrading the performance of the system. This latter problem can be overcome by coupling the detection system to the systems described above for isolating the targets [109]. Another potential approach is to use a specific property of the pathogens that enable them to provide a unique and recognizable signature. One such property that can be exploited is the ability of bacteria to store charge. Direct measurements [110] of the amount of charge in an individual bacte-
290
Sengupta, Gordon, and Chang
rium (~10-13 Coulombs) and their zeta potential (~20 mV) suggest that an individual bacterium, if considered a solid sphere of radius 1 µm, has an effective dielectric constant of ~10,000. This remarkably high number probably arises due to the presence of proteins (such as those forming the various ion channels and regulators) that are extremely efficient in capturing electrons. In contrast, the dielectric constant of aqueous solutions is typically about 80, and so the presence of even a few bacteria should greatly alter the dielectric constant of the solution. Unfortunately, the “double layer” capacitance at the electrode solution interface effectively screens this bacterial capacitance that is dispersed in the bulk solution, especially when operating at frequencies below 1 MHz [111]. However, by modulating the geometry and positioning of the electrodes in a manner that increases the bulk resistance (R), and hence increases the RC time of the medium, the measured reactance (“imaginary” or “out-of-phase” component of the measured impedance) can be made sensitive to bulk capacitance. Consequently signatures of bacterial presence and proliferation at concentrations low enough (100-1000 CFU/ml) to be applicable to many “real world” problems have been detected [112]. Optical techniques Optical detection strategies are arguably the most widely employed platforms for microfluidic analyte detection. Such strategies are diverse, and examples of pathogen microfluidic detection based on absorbance [113], chemiluminescence [114], fluorescence [115, 116], light scattering [117], Raman spectroscopy [118], and refractive index techniques [119] can be found in the literature. A comprehensive treatment of these techniques is beyond the scope of this text, and consequently those readers looking for a rigorous discussion are referred to many excellent reviews [120-126]. Each of the detection techniques listed above have specific advantages, however, fluorescence detection schemes are the most widely employed. Such techniques are commonly utilized in conjunction with microfluidic immunoassays [127, 128], electrophoresis separation schemes [116, 129], and polymerase chain reaction (PCR) amplification of DNA [130, 131]. Fluorescence is popular due to its high sensitivity, allowing for single molecule detection [132, 133], high selectivity, and easy integration of fluorescent tags within the microfluidic system. The joining of fluorescence-based detection platforms with on-chip assay and separation schemes has become prevalent. For instance, Xiang et al. produced an immunoassay device that was capable of detecting E.coli O157:H7 at a limit of detection of 0.3 ng/µl [127]. A membrane-based immunoassay system was developed by Floriano et al. that incorporated fluorescent detection of
Microfluidic Diagnostic Systems
291
Bacillus globigii at a detection limit of as little as 500 spores [128]. Realtime PCR chips that employ fluorescent detection schemes have been developed for the amplification of hepatitis B virus [130] and E. coli O157:H7 stx1 [131]. Electrophoretic separation of PCR amplicons and subsequent fluorescent identification has been demonstrated for the detection of E.coli O157:H7 and Salmonella typhimurium at a starting concentration of as low as 6 copies of target DNA [116]. While fluorescent-based detection schemes are popular within microfluidic diagnostics, the majority of these techniques still rely on relatively bulky external equipment, such as confocal microscopes, preventing them from truly meeting the goals established for an ideal microfluidic diagnostic device. Most commonly, a microscope, connected to either a CCD camera or a photomultiplier tube (PMT), is focused onto the microchip at an appropriate location [125]. Efforts to miniaturize optical detection platforms are currently underway, and are discussed in detail elsewhere [120, 126]. Light emitting diodes (LEDs) coupled with optical filters and silicon photodiodes offer the potential for developing a miniaturized optical detection platform [134]. In producing a hand-held device designed for real-time PCR, Higgins et al developed a miniature optical platform employing two LEDs, at wavelengths of 490 nm and 525 nm, along with a miniaturized PMT [135]. This system rapidly and successfully detected Bacillus anthracis and Erwinia herbicola. Flow cytometry Flow cytometry is an integrated technique for counting, examining, and sorting particles suspended in a fluid. There are three key components of a typical flow cytometer, the first being a fluidic mechanism that causes all of the particles in a suspension to line up in a single file as they flow down a channel. The second key component is a set of detectors (such as lasers of different wavelengths) that can probe individual particles, along with the fluid stream, flowing past the detector and obtain information (such as whether a given cell has taken up a particular fluorescent dye) that indicates one or more specific properties of the cell. This information, along with the known velocity of the particle traveling down the channel, can then be taken advantage of by the third key component of the flow cytometer to steer target particles/cells to specific downstream collection chamber(s), while the remaining particles are discarded. Although conventional state-of the-art flow cytometers can measure and subsequently sort particles based on a combination of as many as ten parameters [136] and/or achieve throughputs as high as ~10,000 cells per second [137], they do suffer from a number of drawbacks. Besides requiring expert operators, they
292
Sengupta, Gordon, and Chang
also require large volumes of sheath fluid (~1 L of sheath fluid per 1 ml of sample) and high performance pumping systems to operate, thereby making them non-portable and prohibitively expensive for routine diagnostic procedures in the clinical setting. Extensive research has been conducted to design and fabricate miniature versions of flow cytometers that replace conventional glass capillary based systems with microfluidic chips that employ integrated optics and hydrodynamic or electrokinetic based flow-switching systems for collecting cells of interest. An excellent review of microfluidic flow cytometry has been provided by Huh et al. [138]. These research efforts, along with others directed towards developing microfluidic pumps, valves and other peripherals [139-141], have already yielded commercial bench-top flow cytometers, such as the Agilent 2100 Bioanalyzer® (Agilent Inc.), the Cyflow® ML (Partec GmbH) and the Mycrocytometer® System (Micronics Inc.). While these systems, like their macroscale counterparts, are primarily designed for mammalian cells, they have also been employed by some researchers to quantify bacteria [142, 143]. This is a difficult task though as bacterial cells are typically 10-100 times smaller than mammalian cells, making them difficult to confine to a single file in a fluid stream, and thus providing unsatisfactory results. Consequently, there have also been efforts to develop cytometers specifically for bacterial cells [143145]. The main drawback of most of these portable microfluidic cytometers (including those under development) is that they typically handle very small volumes of sample (for instance, 10 µl for the 2100 Bioanalyzer), which may not be suitable for operating a real-world sample. Research to increase the throughput of such devices continues, especially for bacterial diagnostics [146]. As is the case for other microfluidic flow-through systems, the likelihood of particulate contaminants in real word samples clogging the device, or otherwise adversely affecting measurements, remains a challenging problem. Mechanical techniques (cantilevers) Cantilevers are beams that are fixed at one end and free at the other. The weight of the beam (when subjected to gravitational forces) and external forces, if any, flexes the beam. The beam’s elasticity, in turn, attempts to restore it to its original shape, thus leading the system to have a characteristic natural frequency of vibration. This frequency, known as the principle resonant frequency, is a function of the effective mass and the spring constant of the beam. Therefore, when the effective mass of the beam in-
Microfluidic Diagnostic Systems
293
creases due to the adhesion of particles (like bacteria) to it, the vibration of the beam is dampened, and the principle resonant frequency decreases [147]. This effect is utilized to detect the presence of bacteria of interest in many microfluidic systems. Typically, antibodies specific to the target pathogen are immobilized on the cantilever, either by physical adsorption [148] or chemically [149], prior to exposing the system to the sample of interest. Ideally, washing with plain buffer removes all particles adhering non-specifically to the cantilever, and the resonant frequency can subsequently be measured. This is typically accomplished through an external instrument, such as an “optical lever” or interferometer, or through changes in the piezoelectric resistance or capacitance of the beam as measured via internally embedded circuit elements [150]. The change in the resonant frequency from its value prior to the introduction of the sample of interest is related to the amount (mass) of bacteria adhering to the cantilever, thereby also allowing for a quick quantification of the target species as well. Cantilevers can be fabricated in silicon based systems using either bulk [151] or surface [152] micromachining. These techniques allow the fabrication of multiple cantilevers within a small surface area [153]. Since different cantilevers can be coated with different antibodies [154], the potential for simultaneous detection of multiple pathogens exists. However, coating multiple antibodies on adjacent micro-cantilevers is a complex task, requiring multiple surface treatment steps (one for each type of antibody being coated/immobilized). Furthermore, this process is made more difficult by problems of “stiction” (fluid remaining stuck underneath the cantilever due to surface tension), that require special treatment procedures like freeze-drying [155], “critical point drying” [156], application of laser pulses [157], etc. after each step involving a different solution. Once these systems are deployed for biosensing, the cantilevers are subjected to multiple stresses that affect their performance. Flow and mixing of the sample solution can give rise to shear forces that bend (and may even break) them. Non-specific adsorption of random species from the sample and electrochemical reactions can change the mass of the cantilever. Moreover, only some of these effects may be corrected for using a reference (control) cantilever in addition to the sensing cantilever. In addition to sharing the problems associated with non-specific adhesion with most antibody-based systems, having their antibodies anchored to a nondispersed surface also requires that the targets have to diffuse to the surface (thereby requiring a long time with solutions having low counts of the target species). Thus, while they have been used in the laboratory to detect
294
Sengupta, Gordon, and Chang
bacteria suspended at reasonably high concentrations (106 / ml and above) in buffers [148, 151], such cantilever based systems may not have the degree of ruggedness that will be required to handle a large number of real world samples.
9.2.2 Pathogen detection/quantification techniques based on detecting metabolites released or consumed When living pathogens respire, they take in sugars, proteins, and other molecules from the environment and release carbon dioxide, pyruvic acid, and large numbers of other compounds. This changes the properties, like pH, dissolved concentration of gasses (oxygen and carbon-dioxide), and electrical conductivity, of the suspending medium. If an appreciable change is detected in these properties, it is taken to imply that pathogens are present that are causing the observed change. A generic growth medium (such as Tryptic Soy Broth, Luria Broth, or Beef Broth) supports the growth of multiple pathogenic and non-pathogenic microbes. Therefore, the use of such media only allows the determination that there exist some viable cells within the sample of interest. That being said, however; selective media can be designed that supports the growth of a limited number of pathogens, typically being those from the same genus. For instance, the Cornell Modified Growth Medium supports the growth of Mycobacteria (a slow growing class of bacteria whose members include the organism causing tuberculosis in humans) while suppressing the growth of other microorganisms [158]. The elapsed time prior to detection (of a specified change in medium properties) can be used to predict the initial pathogen load. This is fairly intuitive given that the greater the initial number of microorganisms in the sample, the higher will be the overall rate of metabolism occurring within the sample, and the faster a detectable change in medium properties will occur. The initial load against elapsed time to detection (TTD) calibration for a particular system (device-medium-organism) can be obtained relatively easily. These methods span a variety of techniques, from those based on detecting the activity of a specific enzyme [159], the release of a specific metabolite like carbon dioxide (radio-labeled [160] or otherwise [161]), looking for an increase in the electrical conductivity [162], a change in the pH (typically observable as a color change within most culturing medium) [163], or finally a change in the oxygen tension of the medium as a whole [164]. A large number of high throughput, but not necessarily “microfluidic” devices rely on this basic principle. These include the Bactec,™ which de-
Microfluidic Diagnostic Systems
295
tects the amount of radio-labeled carbon dioxide released; the ESP® Culture System (TREK Diagnostic Systems), which detects a decrease in oxygen tension; Coli-CheckTM swabs, which use Bromocresol Purple as an indicator to measure a decrease in pH due to bacterial metabolism, and the Bactometer™ (Bactomatic Ltd.), Malthus 2000™ (Malthus Instruments Ltd.), and RABIT™ (Don Whitley Scientific Ltd.) systems that each use electrical impedance. These devices are typically designed to handle samples of ~1-10 ml, and are used for a variety of applications, such as environmental water quality testing, food safety, and veterinary and medical diagnostics. As these devices/methods have greatly simplified handling procedures and incorporated many other features (such as automated readouts) that aid in enabling high-throughout operation, they have acquired broad acceptance in the applied microbiology community. However, the common operating principle, monitoring the sample for effects of microbial metabolism, fundamentally limits how quickly these devices can deliver reliable results. In other words, their “time-to-detection” is still quite high, often being comparable to that of the traditional plating technique. In some cases, such as when the initial bacterial load is very low (~100 CFU/ml or lower), the automated techniques may actually take more time than the traditional plating technique (in which the TTD is independent of bacterial load). This is because the amount of metabolite processed by an individual bacterium is extremely small. Given that the specific oxygen consumption rate for E. coli has been estimated to be ~20 milli-moles of oxygen per hour per gram (dry weight) of bacteria [165], and a typical bacterium has a dry weight of ~10-12 grams [166], a fairly active individual E. coli bacterium consumes ~2 x 10-14 moles of oxygen per hour. A typical well-oxygenated sample will have a dissolved oxygen concentration of ~2 x 10-4 M (2 x 10-7 moles in 1 ml of sample). If the sample has a moderately low initial concentration of bacteria to begin with (~1000 CFU/ml), then the hourly consumption of oxygen for 1000 bacteria (~2 x 10-11 moles) represents a 0.01% change in the concentration of dissolved oxygen. Such a change is so small that it is practically undetectable. Given the exponential growth nature of bacteria, a subsequent change in the concentration of the metabolite being monitored of 2-5% in 8-12 hours for E. coli would be expected. At that point, the presence of the original 1000 could be confirmed. However, in this case, the TTD would be comparable to that obtained with the traditional plating technique. Lower initial loads would further increase this TTD, and slower metabolizing/growing bacteria would further stretch the TTD. Consider that the detection times for mycobacteria are typically 10-45 days using such automated techniques [164].
296
Sengupta, Gordon, and Chang
Many microfluidic counterparts of these above devices essentially apply the same principles to micro-liter (or smaller) volumes rather than the 110ml volumes characteristic of the above devices. Bashir and co-workers have used both impedance [167] and pH [168] as the quantities monitored to register the occurrence of bacterial metabolism. Gas sensors [169] and biosensors for specific chemicals [170] have also been incorporated within microfluidic systems for detecting/monitoring bacterial growth and proliferation. Interestingly, when cells are confined to small volumes (on the order of picoliters) and consequently are in close proximity to the electronic sensing elements, it has also been possible to use thermal sensors [171] to record bacterial metabolism. The TTD of the microfluidic versions of the metabolism based detection systems, however, is limited by the same basic principle that stymies their macrofluidic counterparts. As discussed earlier, if large aliquots need to be sampled due to concerns arising from expectedly low bacterial loads, a pre-concentration step may have to be performed upstream of the reactor. Devices have been developed [86, 172] that employ dielectrophoresis to concentrate bacterial cells from a comparatively larger volume to a smaller one where they are subsequently incubated. In general, as shown in Figure 9.5, there exists an inverse logarithmic relationship between initial bacterial load and time to detection (TTD). This concentration process cuts the TTD by increasing the initial load. Despite the benefits provided by pre-concentration, monitoring bacterial metabolism may not be the most sensitive way to detect the presence of pathogens as evidenced by the high TTDs that are typically encountered. However, this detection platform does have a number of other redeeming features, such as the relative ease with which these devices may be fabricated, assembled, and coupled to automated fluid handling systems and electronic readouts. In addition, there exists the advantage of the recognition within the diagnostic community at large of principles involved within these devices that have been previously utilized in pathogen detection strategies. The Association of Official Analytic Chemists International (AOAC) subjects new techniques to two stages of approval: a “first action” that allows its incorporation in commercial devices under supervision, and a “final action” that approves the method for widespread use after the new method has satisfactorily met its benchmarks in the commercial setting. Impedance based methods for detecting bacterial metabolism in microfluidic systems have been accepted as a “first action” method. It would hence not be surprising to see a commercial microfluidic device of this type in the near future.
Microfluidic Diagnostic Systems
297
Fig. 9.5 Graph showing the inverse logarithmic relationship between initial pathogen load and time to detection (TTD) for systems that detect the presence of pathogens based on a change in the physical properties of the system (such as conductivity or pH) brought about by the pathogens’ metabolism. Reprinted with permission of Wiley-Liss, Inc., a subsidiary of John Wiley & Sons, Inc.: Biotechnology and Bioengineering, vol. 92, issue 6,copyright 2005, [168].
9.2.3 Pathogen detection/quantification through microfluidic immunoassays and nucleic acid based detection platforms Immunoassays and nucleic acid based assays are two vital pathogen detection technologies, and consequently efforts to create miniaturized counterparts are significant. Immunoassays, protein binding assays utilized for the detection of antigens and antibodies, are most commonly carried out as an ELISA. Nucleic acid based assays utilize genetic material, most commonly DNA (and sometimes RNA), for detection and quantification. The fundamental methods developed for nucleic acid diagnostics are enzymatic DNA restriction, nucleic acid hybridization, PCR, and fluorescence-based techniques [173]. Of these techniques, DNA amplification procedures utilizing PCR and PCR-based methods (real-time PCR, strand displacement amplification, nested PCR, etc) are of particular interest given that DNA is present at a low concentration within a typical biological sample. Conse-
298
Sengupta, Gordon, and Chang
quently the genetic identification techniques discussed within this section will focus on microfluidic PCR-based technologies. Several excellent reviews of pathogen identification methods through diverse nucleic acid based assays can be found [173, 174]. 9.2.3.1 Microfluidic immunoassay (ELISA) pathogen identification
Immunoassays are arguably the most heavily utilized quantitative pathogen detection platform [175, 176]. These assays take advantage of the specificity of formation of an antibody-antigen complex, which is a result of the fact that such antibodies are produced based on a specific antigen. Therefore, in use within a diagnostic assay, a given antibody will remain specific to the antigen for which it was developed. As mentioned above, one of the most fundamental immunoassay techniques is the ELISA. There are several different ELISA types, including direct, competitive, and sandwich formats. The sandwich ELISA will likely be one of the primary immunoassay platforms employed on-chip given that it generally exhibits impressive sensitivity and specificity traits, along with favorable kinetics [176]. This assay, as it is traditionally performed, is presented in Figure 9.6.
Fig. 9.6 Basic schematic (minus the rinsing and blocking steps) for a traditional sandwich ELISA. In the first step (A), a capture antibody specific to the target antigen is attached to the surface of a microwell plate. This is followed with a rinsing step, a blocking step, and an additional rinsing step. Sample solution, containing target antigen, is added next, allowing for the target antigen to bind to the capture antibody (B). After an additional rinsing step, a detection antibody, which is generally covalently bound to an enzyme, is introduced and allowed to selectively bind to the captured target antigen (C). Finally, after a last rinse, a substrate molecule is added and converted into a product molecule via the bound enzyme (D). This product molecule is utilized as a means of detecting and quantifying the target antigen.
ELISA assays have become commonplace within diagnostics and have been tailored to a wide variety of detection targets. These assays are designed to detect an antigen directly associated with the pathogen or an an-
Microfluidic Diagnostic Systems
299
tigen/antibody produced within a host system as a result of the invasion of the pathogen. Several examples include the detection of Escherichia coli (E. coli) O157:H7 [113, 177, 178], exotoxins from Salmonella and Clostridium botulinum [179, 180], and Helicobacter pylori [117]. One commercial detection platform based on a sandwich ELISA assay is the VIDAS system marketed by Biomerieux Inc. [178]. This system, which captures bacteria utilizing an antibody specific to an antigen expressed on the surface of the cell, is capable of detecting campylobacter, E. coli O157:H7, Listeria spp, Listeria monocytogenes, Salmonella, Shigella, and Staphylococcal enterotoxins [178]. Another commercial use of the ELISA is within HIV detection. Currently, the majority of HIV testing centers within the United States employ an ELISA that targets antibodies produced as a result of the invasion of the virus [181]. The sensitivity and specificity of this test are both greater than 99% [181]. A positive ELISA result for HIV is typically followed up with a repeat ELISA and then a Western Blot test for final validation. Conventional ELISA assays are typically performed on a multi-well p and are static. Hence the transport of the target molecule to the reactive surface is through diffusion. This process is hampered in terms of being both labor and time intensive, with assay times ranging from at least a few hours to a few days. These drawbacks can often lead to inconsistencies in the obtained assay results and are problematic in producing a rapid diagnostic assay that can be utilized for POC healthcare solutions. The development of microfluidic ELISA platforms will likely relieve these stresses. While microfluidic-based ELISA has not received the same attention that nucleic acid based microfluidic technologies have, the trend is beginning to change [182, 183]. Commercial microfluidic immunoassays are beginning to appear, with one example being the Triage® system developed by Biosite® Inc. (San Diego, CA, USA) to detect several targets, including Clostridium difficile. Microfluidic devices are attractive platforms for immunoassays given that they can alleviate the time and labor difficulties of conventional ELISA as transport is through both diffusion and convection, transport distances are greatly reduced, and the assay can largely (if not completely) be automated. Moreover, such devices have the potential to be fabricated and operated inexpensively, and provide the opportunity to integrate the assay with on-chip sample preparation and target detection, providing a microTotal Analysis System (µ-TAS). The majority of developed microfluidic-based ELISA platforms have either utilized the surface of a microchannel [113, 176, 184-186] or the surface of numerous micro- or nano-beads that are trapped by various means
300
Sengupta, Gordon, and Chang
within a microchannel or a microchamber [187-189]. One of the key issues involved in localizing an immunoassay to the surface of a microchannel wall is the efficient immobilization of antibody (or protein) to the surface while preventing non-specific protein adsorption that would both interfere with the assay and also reduce the available surface area for target protein immobilization. This issue was discussed in a review written by Lee [190] and has been considered in several publications [176, 185, 191]. In developing a microfluidic immunoassay device for the detection of E. coli O157:H7, Bai et al. functionalized a poly(methyl methacrylate) (PMMA) surface with poly(ethyleneimine) (PEI) to introduce amine groups to the polymer channel walls [191]. This was followed by glutaraldehyde addition to provide aldehyde groups for protein binding. An affinity purified antibody to E. coli O157:H7 cells was then allowed to adsorb to the surface of the channel via the aldehyde groups, followed by the introduction of a sample suspension containing inactivated cells which was allowed to bind to the immobilized antibody. Finally, peroxidase-labeled affinity purified antibody to the cells was added. The peroxidase enzyme catalyzed a substrate to product conversion, producing fluorescence for detection purposes. This surface treatment method improved antibody binding, and compared to untreated PMMA microchannels, an approximately 45 times greater signal and 3 times greater signal/noise ratio was achieved. Additionally, this microchannel device required only 2 minutes for E. coli O157:H7 capture and required a mere 5-8 cells within the initial sample aliquot, both of which are an improvement compared to traditional multiwell plate ELISA. The use of beads within an immunoassay microfluidic device increases the complexity of the device and requires methods for trapping and handling the beads on-chip [192]. However, the use of beads is becoming more commonplace [187-189, 193] because of inherent advantages, such as an increased available surface area to volume ratio (which in-turn increases the real-estate that can be taken advantage of for protein binding), further reduced antigen-antibody molecular diffusion distance, potential for multiplexing, ease of manipulation, and ease of performing surface chemistry modifications. One specific example of a bead-based microfluidic immunoassay platform was provided by Liu et al. for the detection of a marine iridovirus [187]. In their work, antibody-coated microbeads (~2.89 µm in diameter) were trapped within a microfluidic channel via an included filter assembly. After incubation with the iridovirus, detection was provided through the insertion of quantum dots attached to the detection antibody. In comparison to a traditional ELISA scheme, this approach was able to detect the marine iridovirus at a significantly improved limit of detection
Microfluidic Diagnostic Systems
301
(22 ng/ml as opposed to 360 ng/ml), in a shorter time period (<0.5 hours as opposed to >3.25 hours), and with less demand on the availability of antibodies (0.035 µg used as opposed to 0.5 µg). 9.2.3.2 Microfluidic PCR based pathogen identification
The identification of pathogens through genetic analysis techniques is becoming more conventional. Several examples include the detection of Helicobacter pylori [194], Mycobacterium tuberculosis [195-197], Neisseria gonnorhoeae [198], Streptococci (Group B) [199], and E. coli O157:H7 [200]. The genetic identification of pathogens is dependent upon the presence of sufficient genetic material (i.e. DNA or RNA). However, it is typical that the target pathogen is present at a dilute concentration, complicating detection. To alleviate this difficulty, DNA amplification becomes essential. This is typically achieved through implementation of the PCR, which incorporates three cycled reaction steps: a denaturing step run at 92-96ºC, an annealing step run at about 40-65ºC, and finally an extension step run at about 72ºC. PCR-based methods for pathogen detection have gained popularity over traditional culturing procedures, primarily due to enhanced assay speed and reliability improvements. The same push for improved assay speed and reliability are now driving the development of microfluidic PCR-based pathogen detection technologies, with the ultimate goal being the development of POC diagnostic devices that meet the goals laid out earlier. Several of the advantages offered in miniaturizing the PCR process are similar to those expected with the miniaturization of immunoassays and include a reduction in sample and reagent volumes and a corresponding reduction in reagent costs, increased portability and disposability, and integration with upstream sample preparation and downstream analysis techniques to provide the highly desired µ-TAS. One notable advantage specific to microfluidic-based PCR is a decreased time of amplification as a result of low thermal capacities and large heat transfer rates (because of large surface area to volume ratios) [201, 202]. Though the miniaturization of PCR offers several advantages, the first micro-PCR device was not described until 1993 [203]. Currently, microfluidic chip based PCR can operate with reaction volumes of less than 200 nl [204-208], and can offer reductions in reaction time through improvements in the on-chip heating and cooling rates. A conventional thermal cycler provides heating and cooling rates of approximately 1-2ºC/s, however, a heating rate of 175ºC/s and a cooling rate of 125ºC/s has now been reported within a microfluidic design [209]. Commercial microfluidic based PCR devices have arisen as
302
Sengupta, Gordon, and Chang
well, and have been employed in detecting a variety of pathogens. For instance, Mycobacterium bovis was detected using the GeneXpert® system produced by Cepheid (Sunnyvale, CA) [210]. Within this system, a filter is employed to capture organisms from a clinical sample, which is then followed by cell lysis. Real-time PCR of the released nucleic acids can then be conducted for pathogen identification. This system is of particular interest as all of these steps, organism isolation, cell lysis, and real-time PCR detection are performed on a single, disposable cartridge. Another example of a commercial microfluidic based PCR device is the BioMark™ 48.48 Dynamic Array developed by the Fluidigm® Corporation (San Francisco, CA). This system is capable of running up to 2304 reactions per chip while only 96 total liquid loading steps are required. The advantages offered by this microfluidic chip, especially the large number of reactions that can be run in parallel, were beneficial to a study aimed at probing individual bacteria within the lignocellulose-decomposing microbial community that exists within wood-feeding termites [211]. Within the literature, there exist several applications of microfluidic PCR amplification of specific pathogens. Some of these examples include; Dengue II virus [212], E. coli SK [213], Salmonella typhimurium [116, 214], severe acute respiratory syndrome (SARS) [215, 216], BK virus [217], Campylobacter jejuni [218], Human Papilloma virus (HPV) [219], HIV [220], Mycobacterium tuberculosis [221], Influenza viral strain A/LA/1/87 [222], Neisseria gonnorhoeae [223], Bacillus anthracis [224], Hepatitis B [225], and Hepatitis C [226]. Another example, E. coli O157:H7, is a food-born pathogen that has been identified as the culprit in several severe human infection outbreaks [227]. This pathogen is particularly dangerous given that a minimal concentration is required for human infection (as low as 10-100 CFU/g) [228], and that the organism produces many clinical manifestations [229]. It has been estimated that there are 73000 cases and 61 deaths per year within the United States alone that are attributable to this pathogen [230]. Traditional detection protocols, namely identifying the bacteria in stool cultures, can be time-intensive [230, 231], requiring at least a 16 hour culture period [232]. In view of this limitation, Koh et al. have described a microfluidic device created from poly(cyclic olefin) for the amplification of E. coli O157:H7 and subsequent on-chip detection through electrophoresis and fluorescence identification of amplicons [116]. The static amplification was carried out within a gel-valve isolated reaction chamber with volumes as small as 29 nL using E. coli O157:H7 serotype specific primers. Heating was accomplished using a localized printed ink-based heater that allowed for a heating rate of ~12ºC/s. A detection limit of approximately 6 DNA copies present within the initial
Microfluidic Diagnostic Systems
303
solution fed to the PCR was determined. Several additional examples of on-chip PCR of E. coli O157:H7 can be found within the literature [233, 234]. There are several issues that complicate the rapid, microfluidic genomic detection of pathogenic substances. These concerns include low concentrations of biological targets and a typically complex biological sample matrix. Taken in context, there is thus some balance between minimum sample volume, rapid assay time, and detection sensitivity that must be considered in practical device fabrication. Considering this, Yang et al. amplified a DNA segment on-chip directly from an E. coli K12 strain (often referred to as colony PCR) [232]. The polycarbonate device employed a serpentine PCR channel for continuous amplification of volumes as large as 40 µl, and target DNA was provided through thermal lysis of the E. coli cells. The amplification was carried out using these cells both within and without the company of 2% sheep blood, and amplification from as few as 10 cells was observed. To block potential PCR inhibitors present within the biological sample matrix, the PCR mixture was supplemented with a specific buffer (Q buffer). Another potential means around this is to isolate the target genomic DNA after on-chip lysis. This was accomplished after thermal lysis of E. coli cells through capture onto probe tagged magnetic particles by Yeung et al. [235]. After DNA isolation, asymmetric PCR was performed to produce amplified single strand DNA that was subsequently detected through electrochemical methods. Other examples of microfluidic amplification of target DNA directly from a biological sample can be found within the literature [211, 217, 218, 236-238]. There remain many key considerations regarding the miniaturization of PCR that must be resolved. Some of these platform design criteria can be seen in comparing the devices described by Koh et al. [116] and Yeung et al. [235]. This includes the PCR format, which can be static with the PCR mixture held within a reaction chamber while the chamber temperature is cycled through the normal temperature ranges, or continuous with the PCR mixture transported through a channel and different regions of the channel held to appropriate temperatures. Each of these designs present different challenges, such as achieving a rapid and precise temperature control and fluid manipulation. Additional considerations include providing for onchip heating and cooling (typically through either contact or non-contact methods), preventing sample evaporation (which is an increasing problem with a reduction in sample volume), ensuring sufficient sample volume to provide enough starting target DNA, preventing cross-contamination, preventing sample adsorption to surface walls, integrating sample preparation
304
Sengupta, Gordon, and Chang
and detection platforms with the PCR on-chip, the material(s) employed for device fabrication, and providing for real-time PCR. As would be expected, the various materials available for construction of microfluidic devices each have different properties that will impact some aspect of the PCR reaction. Thus far, most microfluidic PCR devices have been fabricated utilizing either silicon or glass [116, 202, 239]. Silicon is advantageous for its high thermal conductivity and glass is advantageous because it lends itself well to integration with optical detection schemes. However, the use of these two materials presents specific challenges. For instance, bare silicon is opaque, complicating the integration of optical detection methodologies, and it is known to inhibit the PCR reaction through sample adsorption to the surface wall [239]. As is also the case with silicon chips, glass carries with it a high fabrication cost, preventing the realization of a truly disposable chip. This lack of disposability increases the risk of crosscontamination through multiple uses [226]. A potential means of accounting for cross-contamination involves control of the micro-chamber surfaces. To accomplish this end, Prakash et al. silanized a glass microchamber surface prior to carrying out the PCR reaction [220]. Following the reaction, the silanized surface of the microchamber was removed and a fresh silanization was carried out. In this manner, cross-contamination was controlled by ensuring sample antigen from a previous amplification did not remain on the wall of the microchamber during future amplification reactions. Another attractive solution to the cross-contamination problem involves the use of certain polymer substrates in the construction of microfluidic chips. Polymers, such as polydimethylsiloxane (PDMS), are relatively inexpensive, potentially allowing for the realization of a truly disposable chip, which would obviously circumvent the problem of crosscontamination. PDMS is also an attractive material as it provides good optical properties and exhibits limited PCR sample adsorption. Consequently there are several examples of the use of PDMS within microfluidic PCR designs [202, 213, 217, 239]. However, the use of PDMS is not without its own difficulties. One such difficulty is that PDMS is permeable, providing the possibility of sample losses. This problem becomes more significant as sample volume is reduced within a microfluidic chip. To avert this obstacle, the use of a vapor barrier has been suggested [240]. In one report, a polyethylene vapor barrier layer was implanted within the PDMS from which the device was fabricated [240]. Amplification was then carried out on 1.75 µl of PCR solution, with yeast genomic DNA as the target. The use of the vapor barrier was shown to reduce fluid loss by at least 3-fold, significantly enhancing the amplification. Finally it should not go without mentioning that sample evaporation is not a problem solely confined to polymer-fabricated devices, as this issue must also be considered when re-
Microfluidic Diagnostic Systems
305
ducing sample volume within silicone and glass devices. Steps taken to minimize this sample loss include the use of mineral oil [209] and microvalves [241]. Recently, valve-less strategies based on diffusion-limited evaporation mechanisms for reducing sample evaporation were shown to be effective by Wang et al. [242]. To accomplish this aim, long and narrow channels were built into the microchip. As convective airflow is not present within these channels, sample evaporation then becomes limited by the lengthy diffusion time of vapor traversing the long, narrow channel. To supplement this strategy, efforts to decrease the liquid evaporation driving force included thermal isolation and vapor replenishment utilizing water. The above discussion highlights only a few of the design issues facing the development of microfluidic PCR devices, and the reader is thus referred to several informative reviews that discuss this topic and also provide descriptions of the current state of microfluidic PCR technology [183, 201, 202, 239, 243-245]. Ultimately, it is unlikely that one particular design and design material will suffice for all potential uses of microfluidic PCR devices. It is expected that the final decision of which materials and designs to employ will be dependent on the specific intended use of the chip and the associated integrated components.
9.3 Future Research Directions The goal of the research described in this chapter is development of more effective microfluidic diagnostic devices. The research generally proceeds along two inter-related fronts that can be viewed as being analogous to the concepts of hardware and software. The “hardware” aspect of this research is directed at designing better micro-devices (like pumps, sensors, etc.), formulating methods to fabricate these micro-devices in or on materials that are biocompatible (or desirable from some other standpoint), and in devising ways to package these systems and interface them with the outside (macro) world. The “software” aspect of this research is directed towards arriving at newer and more effective strategies for isolating and/or detecting target pathogens. Thus far, most groups have been focused on different areas of "hardware" and "software" research, such as either developing improved biosensors for pathogen detection or developing new processes for the isolation and/or selective amplification of targets. While these efforts have produced very interesting and encouraging results, the development of a practical and ro-
306
Sengupta, Gordon, and Chang
bust device that fulfils the pathogen detection goals laid out in the introduction of this chapter will likely require considerable focus on combining these varied approaches into an integrated microfluidic system. Such a system will conceivably (at least when describing first generation devices) employ a short pre-culture step in order to increase the concentration of detection targets. This step will be followed by relatively simple separation steps (perhaps utilizing phenomena such as DEP), and then the device will end with a sensitive biosensor of some type. In designing this system, it will be important to consider that the optimization of target pathogen detection (i.e. detecting low number of target pathogens with high specificity) might require the operation of the initial device stages (selective amplification and separation) sub-optimally. The development of effective microfluidic devices for pathogen detection will also have to respond to new challenges that will arise from an increased understanding of the role microbes play within disease. For instance, recent work [246] has uncovered links between periodontal disease and the infestation of gums by archaea (a class of microscopic organism that is believed to have evolved separately from bacteria and other cellular organisms). Periodontal disease, in turn, has been identified as a contributing factor to endocarditis, atherosclerosis, stroke, and preterm delivery of low-birth-weight infants. Unfortunately, little is currently known about the exact mechanisms by which archaea could lead to these complications. Additionally, the protocols for quantifying the presence of such microscale organisms have either not been standardized, or are known only to a few expert researchers. Consequently, diagnostic devices (both micro- and macro-scale) that can successfully detect such microbes have yet to be developed. Furthermore, diagnostic devices have largely been developed according to the conventional medical paradigm of “one disease, one pathogen.” This model states that if a patient’s symptoms match those of a particular disease, then the symptoms are expected to be caused by one type of pathogen (or a few related types) that have been previously associated with the respective disease in question. However, recent research has suggested that microbial “communities” may in fact either be responsible for and/or contributing towards several aggravating conditions, such as obesity [247] and Inflammatory Bowel Disease (IBD) [248, 249]. In such instances, the relative number, pathogen-host and pathogen-pathogen metabolic interaction, and spatial distribution of the micro-organisms likely determines their resulting impact (regarding both what the impact is and its severity) on the host. All these factors will add up to alter the technical goals required of
Microfluidic Diagnostic Systems
307
microfluidic diagnostic devices and place increased demands on their capabilities in the future.
References 1. Feldmann H, Czub M, Jones S, Dick D, et al. (2002) Emerging and re-emerging infectious diseases. Medical Microbiology and Immunology 191:63-74. 2. Morens DM, Folkers GK, Fauci AS (2004) The challenge of emerging and reemerging infectious diseases. Nature 430:242-249. 3. Johnson N, Mueller J (2002) Updating the accounts: global mortality of the 1918-1920 "Spanish" influenza pandemic. Bulletin of the History of Medicine 76:105-115. 4. (2000) Guidelines for surveillance, prevention, and control of West Nile virus infection - United States (Reprinted from MMWR, vol 49, pg 25, 2000). Jama-Journal of the American Medical Association 283:997-998. 5. Fauci AS (2001) Infectious diseases: Considerations for the 21st century. Clinical Infectious Diseases 32:675-685. 6. Jernigan DB (2003) Investigation of bioterrorism-related anthrax, United States, 2001: Epidemiologic findings (vol 8, pg 1019, 2001). Emerging Infectious Diseases 9:140-140. 7. Cadieux G, Tamblyn R, Dauphinee D, Libman M (2007) Predictors of inappropriate antibiotic prescribing among primary care physicians. Canadian Medical Association Journal 177:877-883. 8. Austin DJ, Kristinsson KG, Anderson RM (1999) The relationship between the volume of antimicrobial consumption in human communities and the frequency of resistance. Proceedings of the National Academy of Sciences of the United States of America 96:1152-1156. 9. Guerrant RL, Blackwood BL (1999) Threats to global health and survival: The growing crises of tropical infectious diseases - Our "unfinished agenda". Clinical Infectious Diseases 28:966-986. 10. Yager P, Edwards T, Fu E, Helton K, et al. (2006) Microfluidic diagnostic technologies for global public health. Nature 442:412-418. 11. Angus D, Linde-Zwirble W, Lidicker J, Clermont G, et al. (2001) Epidemiology of severe sepsis in the United States: analysis of incidence, outcome, and associated costs of care. Critical Care Medicine 29:1303-1310. 12. Thom T, Haase N, Rosamond W, Howard VJ, et al. (2006) Heart Disease and Stroke Statistics--2006 Update. A Report From the American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Circulation 105:171600. 13. U.S. Cancer Statistics Working Group (2005) United States Cancer Statistics: 1999-2002 Incidence and Mortality Web–based Report Version. Atlanta. U.S. Department of Health and Human Services, Centers for Disease Control and Prevention and National Cancer Institute.14. Friedman G, Silva E, Vincent J-
308
Sengupta, Gordon, and Chang
L (1998) Has the mortality of septic shock changed over time ? Critical Care Medicine 26:2078-2086. 15. Bhattacharya S (2005) Blood culture in India: A proposal for a national programme for early detection of sepsis. Indian Journal of Medical Microbiology 23:220-226. 16. Doern GV, Vautour R, Gaudet M, Levy B (1994) Clinical impact of rapid invitro susceptibility testing and bacterial identification. Journal of Clinical Microbiology 32:1757-1762. 17. Gómez J, Simarro E, Baños V, Requena L, et al. (1999) Six-Year Prospective Study of Risk and Prognostic Factors in Patients with Nosocomial Sepsis Caused by Acinetobacter baumannii. European Journal of Clinical Microbiology & Infectious Diseases 18:358-361. 18. Kang C-I, Kim S-H, Kim H-B, Park S-W, et al. (2003) Pseudomonas aeruginosa Bacteremia: Risk Factors for Mortality and Influence of Delayed Receipt of Effective Antimicrobial Therapy on Clinical Outcome. Clinical Infectious Diseases 37:745-751. 19. Stanier RY, Ingraham JL, Wheelis ML, Painter PR, The Microbial world, Prentice-Hall, Englewood Cliffs, NJ 1986. 20. Eckberg PB, Bik EM, Bernstein CN, Purdom E, et al. (2005) Diversity of the Human Intestinal Microbial Flora. Science 308:1635-1638. 21. Lowry SP, Melnick JL, Rawls WE (1971) Investigation of Plaque Formation in Chick Embryo Cells as a Biological Marker for Distinguishing Herpes Virus Type 2 from Type 1. Journal of General Virology 10:1-9. 22. Yagupsky P, Nolte FS (1990) Quantitative Aspects of Septicemia. Clinical Microbiology Reviews 3:269-279. 23. Padhye NV, Doyle MP (1991) Production and characterization of a monoclonal antibody specific for enterohemorrhagic Escherichia coli of serotypes O157:H7 and O26:H11. Journal of Clinical Microbiology 29:99-103. 24. Jonsson U, Malmqvist M, Ronnberg I (1985) Immobilization of immunoglobulins on silica surfaces. Biochemical Journal 227:363-371. 25. Holt DB, Gauger PR, Kusterbeck AW, Ligler FS (2002) Fabrication of a capillary immunosensor in polymethyl methacrylate Biosensors and Bioelectronics 17:95-103. 26. Kato K, Ikada Y (1996) Immobilization of DNA onto a polymer support and its potentiality as immunoadsorbent Biotechnology and Bioengineering 51:581-590. 27. Safranj A, Kiaei D, Hoffman AS (1991) Antibody Immobilization onto Glow Discharge Treated Polymers. Biotechnology Progress 7:173-177. 28. Park I-S, Kim N (1998) Thiolated Salmonella antibody immobilization onto the gold surface of piezoelectric quartz crystal. Biosensors and Bioelectronics 13:1091-1097. 29. Atlas R, Williams J, Huntington. MK (1995) Legionella contamination of dental-unit waters. Applied and Environmental Microbiology 61:1208-1213. 30. Forbes BA, Sahm DF, Weissfeld AS, Diagnostic Microbiology, Mosby, St. Louis 1998.
Microfluidic Diagnostic Systems
309
31. Zanzotto A, Szita N, Boccazzi P, Lessard P, et al. (2004) Membrane-aerated microbioreactor for high-throughput bioprocessing. Biotechnology and Bioengineering 87:243-254. 32. Huang C-W, Lee G-B (2007) A microfluidic system for automatic cell culture. JOURNAL OF MICROMECHANICS AND MICROENGINEERING 17:1266-1274. 33. Steinhaus B, Garcia ML, Shen AQ, Angenent LT (2007) A Portable Anaerobic Microbioreactor Reveals Optimum Growth Conditions for the Methanogen Methanosaeta concilii Applied and Environmental Microbiology 73:16531658. 34. Zhang Z, Boccazzi P, Choi H-G, Perozziello G, et al. (2006) Microchemostat microbial continuous culture in a polymer based, instrumented microbioreactor. Lab on a Chip 6:906-913. 35. Merkel T, Bondar V, Nagai K, Freeman B, Pinnau I (2000) Gas sorption, diffusion and permeation in poly(dimethylsiloxane). Journal of Polymer Science; Part B - Polymer Physics 38:415-434. 36. Zanzotto A, Szita N, Schmidt MA, Jensen KF, 2nd Annual International IEEE-EMBS Special Topic Conference on Microtechnologies in Medicine and Biology, Madison, WI 2002, pp. 164-168. 37. Chang W-J, Akin D, Sedlek M, Ladisch M, Bashir R (2003) Poly(dimethylsiloxane) (PDMS) and Silicon Hybrid Biochip for Bacterial Culture. Biomedical Microdevices 5:281-290. 38. Pasmore M, Department of Chemical Engineering, University of Colorado, Boulder, CO 1998, p. 207. 39. Larsen DH, Dimmick RL (1964) Attachment and growth of bacteria on surfaces of continuous culture vessels. Journal of Bacteriology 88:1380-1387. 40. Balagadde FK, You L, Hansen CL, Arnold FH, Quake SR (2005) Long-Term Monitoring of Bacteria Undergoing Programmed Population Control in a Microchemostat. Science 309:137-140. 41. Sengupta S, Chang H-C, in: Li, D. (Ed.), Encyclopedia of Microfluidics and Nanofluidics, Springer-Verlag In Print. 42. Andersson H, Van-der-Wijngaart W, Enoksson P, Stemme G (2000) Micromachined flow-through filter-chamber for chemical reactions on beads. Sensors and Actuators B: Chemical 67:203-208. 43. Kim DS, Lee SH, Ahn CH, Leed JY, Kwon TH (2006) Disposable integrated microfluidic biochip for blood typing by plastic microinjection moulding. Lab on a Chip 6:794-802. 44. Gatimu EN, King TL, Sweedler JV, Bohn PW (2007) Three-dimensional integrated microfluidic architectures enabled through electrically switchable nanocapillary array membranes. Biomicrofluidics 1. 45. Metz S, Trautmann C, Bertsch A, Renaud P (2004) Polyimide microfluidic devices with integrated nanoporous filtration areas manufactured by micromachining and ion track technology. Journal of Micromechanics and Microengineering 14:324-331. 46. Moorthy J, Beebe DJ (2003) In-situ fabricated porous filters for microsystems. Lab on a Chip 3:62-66.
310
Sengupta, Gordon, and Chang
47. Wang P, Chen Z, Chang H-C (2006) A new electro-osmotic pump based on silica monoliths. Sensors and Actuators B: Chemical 113:500-509. 48. Chen Z, Wang P, Chang H-C (2005) An electro-osmotic micropump based on monolithic silica for micro-flow analyses and electro-sprays. Analytical and Bioanalytical Chemistry 382:817. 49. Gordon J, Senapati S, Hou D, Nowak C, et al., American Institute of Chemical Engineers (AIChE) Annual Meeting, Salt Lake City, UT 2007. 50. Zhu L, Zhang Q, Feng H, Ang S, et al. (2004) Filter based microfluidic device as a platform for immunofluorescent assay of microbial cells. Lab on a Chip 4:337-341. 51. Peh XL, Zhu L, Teo CY, Ji HM, et al., The 14th International Conference on Solid State Sensors, Actuators and Microsystems, 1861-1864, Lyon, France 2007. 52. Sengupta S, Wang P, Gagnon Z, Chang H-C, American Institute of Chemical Engineers (AIChE) Annual Meeting, San Francisco, CA 2006. 53. Hatch GP, Stelter RE (2001) Magnetic design considerations for devices and particles used for biological high-gradient magnetic separation (HGMS) systems. Journal of Magnetism and Magnetic Materials 225:262-276. 54. Fisher MD, Frost SC (1997) Using magnetic beads to isolate inside-out glut1containing vesicles from 3T3-L1 adipocyte plasma membranes. Analytical Biochemistry 251:125-126. 55. Han KH, Frazier AB (2004) Continuous magnetophoretic separation of blood cells in microdevice format. Journal of Applied Physics 96:5797-5802. 56. Liu YJ, Guo SS, Zhang ZL, Huang WH, et al. (2007) A micropillar-integrated smart microfluidic device for specific capture and sorting of cells. Electrophoresis 28:4713-4722. 57. Cho YK, Lee JG, Park JM, Lee BS, et al. (2007) One-step pathogen specific DNA extraction from whole blood on a centrifugal microfluidic device. Lab on a Chip 7:565-573. 58. Lien KY, Lee WC, Lei HY, Lee GB (2007) Integrated reverse transcription polymerase chain reaction systems for virus detection. Biosensors & Bioelectronics 22:1739-1748. 59. Xia N, Hunt TP, Mayers BT, Alsberg E, et al. (2006) Combined microfluidicmicromagnetic separation of living cells in continuous flow. Biomedical Microdevices 8:299-308. 60. Pamme N (2006) Magnetism and microfluidics. Lab on a Chip 6:24-38. 61. Smistrup K, Hansen O, Bruus H, Hansen MF (2005) Magnetic separation in microfluidic systems using microfabricated electromagnets-experiments and simulations. Journal of Magnetism and Magnetic Materials 293:597-604. 62. Pekas N, Granger M, Tondra M, Popple A, Porter MD (2005) Magnetic particle diverter in an integrated microfluidic format. Journal of Magnetism and Magnetic Materials 293:584-588. 63. Mirowski E, Moreland J, Russek SE, Donahue MJ (2004) Integrated microfluidic isolation platform for magnetic particle manipulation in biological systems. Applied Physics Letters 84:1786-1788.
Microfluidic Diagnostic Systems
311
64. Pankhurst QA, Connolly J, Jones SK, Dobson J (2003) Applications of magnetic nanoparticles in biomedicine. Journal of Physics D-Applied Physics 36:R167-R181. 65. Gijs MAM (2004) Magnetic bead handling on-chip: new opportunities for analytical applications. Microfluidics and Nanofluidics 1:22-40. 66. Giddings JC (1989) Harnessing electrical forces for separation. Capillary zone electrophoresis, isoelectric focusing, field-flow fractionation, split-flow thincell continuous-separation and other techniques. Journal of Chromatography 480:21-33. 67. Sengupta S, Todd P, Thomas N (2002) Multistage Electrophoresis II: Treatment of a kinetic separation as a pseudo-equilibrium process. Electrophoresis 23:2064-2073. 68. Harden VP, Harris JO (1953) The Isoelectric Point of bacterial Cells. Journal of Bacteriology 65:198-202. 69. Dyar M, Ordal E (1946) Electrokinetic Studies on Bacterial Surfaces: I. The Effects of Surface-active Agents on the Electrophoretic Mobilities of Bacteria. Journal of Bacteriology 51:149-167. 70. Bauer J, Cell Electrophoresis, CRC Press, Boca Raton, FL 1994. 71. Végvári Á, Hjertén S (2003) Hybrid microdevice electrophoresis of peptides, proteins, DNA, viruses, and bacteria in various separation media, using UVdetection Electrophoresis 24:3815 - 3820. 72. Kremser L, Blaas D, Kenndler E (2004) Capillary electrophoresis of biological particles: Viruses, bacteria, and eukaryotic cells. Electrophoresis 25:22822291. 73. Poe BG, Arriaga EA, in: Landers, J. (Ed.), Handbook of Capillary Electrophoresis (3rd Edition), CRC Press, Boca Raton, FL 2007. 74. Shintania T, Yamada K, Torimura M (2002) Optimization of a rapid and sensitive identification system for Salmonella enteritidis by capillary electrophoresis with laser-induced fluorescence. FEMS Microbiology Letters 210:245249. 75. Armstrong DW, Schneiderheinze JM, Kullman JP, He L (2001) Rapid CE Microbial Assays for Consumer Products That Contain Active Bacteria. FEMS Microbiology Letters 194:33-37. 76. Moon BG, Lee Y-I, Kang SH, Kim Y (2003) Capillary Electrophoresis of Microbes. Bulletin of the Korean Chemical Society 24. 77. Armstrong DW, Schulte G, Schneiderheinze JM, Westenberg DJ (1999) Separating Microbes in the Manner of Molecules: I. Capillary Electrokinetic Approaches. Analytical Chemistry 71:5465-5469. 78. Mutlu S, Yu C, Selvaganapathy P, Svec F, et al., IEEE MEMS 2002 Las Vegas, NV 2002, pp. 19-24. 79. Takhistov P, Duginova K, Chang H-C (2003) Electrokinetic Mixing Vortices due to Electrolyte Depletion at Microchannel Junctions. Journal of Colloid and Interfacial Science 263:133-143. 80. Gonzalez A, Ramos A, Green NG, Castellanos A, Morgan H (2000) Fluid Flow Induced by Non-Uniform AC Electrolytes on Micro-Electrodes: A Linear Double Layer Analysis. Physical Review E 61.
312
Sengupta, Gordon, and Chang
81. Ajdari A (2000) Pumping Liquids Using Asymmetric Electrode Arrays. Physical Review E 61:R45-R48. 82. Pohl HA, Dielectrophoresis, Cambridge University Press, Cambridge 1978. 83. Jönsson M, Aldaeus F, Johansson LE, Lindberg U, et al., Micro System Workshop '04, Ystad, Sweden 2004. 84. Markx GH, Huang Y, Zhou XF, Pethig R (1994) Dielectrophoretic characterization and separation of microorganisms. Microbiology 140:585-591. 85. Lapizco-Encinas BH, Simmons BA, Cummings EB, Fintschenko Y (2004) Insulator-based dielectrophoresis for the selective concentration and separation of live bacteria in water. Electrophoresis 25:1695-1704. 86. Li H, Bashir R (2002) Dielectrophoretic separation and manipulation of live and heat-treated cells of Listeria on microfabricated devices with interdigitated electrodes. Sensors and Actuators B 86:215-221. 87. Visuri SR, Ness K, Dzenitis J, Benett B, et al., 2nd Annual International IEEEEMB Special Topic Conference on Microtechnologies in Medicine & Biology 2003. 88. Gascoyne PRC, Vykoukal J (2003) Particle separation by dielectrophoresis Electrophoresis 23:1973-1983. 89. Gordon JE, Gagnon Z, Chang H-C (In press) Dielectrophoretic discrimination of bovine red blood cell starvation age by buffer selection and membrane cross-linking. Biomicrofluidics. 90. Gagnon Z, Gordon JE, Chang H-C (In press) Bovine red blood cell starvation age discrimination through an optimized dielectrophoretic approach with buffer selection and membrane cross-linking. Electrophoresis. 91. Wu J, Ben Y, Battigelli D, Chang H-C (2005) Long-range AC Electrokinetic Trapping and Detection of Bioparticles. Industrial and Engineering Chemistry Research 44:2815-2822. 92. Gagnon Z, Chang H-C (2005) Aligning fast alternating current electroosmotic flow fields and characteristic frequencies with dielectrophoretic traps to achieve rapid bacteria detection. Electrophoresis 26:3725-3737. 93. Hou D, Maheshwari S, Chang H-C (2007) Rapid bioparticle concentration and detection by combining a discharge driven vortex with surface enhanced Raman scattering. Biomicrofluidics 1. 94. Hou D, Chang H-C (2006) Electrokinetic particle aggregation patterns in microvortices due to particle-field interaction. Physics of Fluids 18. 95. Cheng I-F, Chang H-C, Hou D, Chang H-C (2007) An integrated highthroughput dielectrophoretic chip for continuous bio-particle filtering, focusing, sorting, trapping, and detection. Biomicrofluidics 1. 96. Wang L, Flanagan LA, Jeon NL, Monuki E, Lee AP (2007) Dielectrophoresis switching with vertical sidewall electrodes for microfluidic flow cytometry. Lab on a Chip 7:1114-1120. 97. Muller T, Gradl G, Howitz S, Shirley S, et al. (1999) A 3-D microelectrode system for handling and caging single cells and particles. Bioesensors and Bioelectronics 14.
Microfluidic Diagnostic Systems
313
98. James CD, Okandan M, Galambos P, Mani SS, et al. (2006) Surface micromachined dielectrophoretic gates for the front-end device of a biodetection system. Transactions of the ASME 128:14-19. 99. Morgan H, Green NG, AC Electrokinetics: colloids and nanoparticles, Research Studies Press, Ltd., Baldock, Hertfordshire, England 2003. 100. Bennett DJ, Khusid B, James CD, Galambos P, et al. (2003) Combined fieldinduced dielectrophoresis and pahse separation for manipulating particles in microfluidics. . Applied Physics Letters 83:4866-4868. 101. Fielder S, Shirley SG, Schnelle T, Fuhr G (1998) Dielectrophoretic sorting of particles and cells in a microsystem. . Analytical Chemistry 70:1909-1915. 102. Holmes D, Morgan H, Green NG (2006) High throughput particle analysis: Combing dielectrophoretic particle focussing with confocal optical detection. . Bioesensors and Bioelectronics 21:1621-1630. 103. Zhou R, Wang P, Chang H-C (2006) Bacteria capture, concentration and detection by AC dielectrophoresis and self-assembly of dispersed single wall carbon nanotubes. Electrophoresis 27:1375. 104. Arumugam PA, Chen H, Cassell AM, Li J (2007) Dielectrophoretic trapping of single bacteria ar carbon nanofiber nanoelectrode arrays. Journal of Physical Chemistry A. 105. Gale BK, Frazier AB, SPIE Symp. Micromach. Microfab.: Micro Fluidic Dev. Syst., Santa Clara, CA 1999, pp. 190-201. 106. Chang B-W, Chen CH, Ding S-J, Chen DC-H, Chang H-C (2005) Impedimetric monitoring of cell attachment on interdigitated microelectrodes. Sensors and Actuators B: Chemical 105:159-163. 107. Suehiro J, Ohtsubo A, Hatano T, Hara M (2006) Selective stection of bacteria by a dielectrophoretic impedance measurement method using an antibodyimmobilized electrode chip. Sensors and Actuators B: Chemical. 108. Javanmard M, Talasaz AH, Nemat-Gorgani M, Pease F, et al. (2007) Targeted cell detection based on microchannel gating. Biomicrofluidics 1. 109. Patolsky F, Leiber C (2004) Electrical Detection of Single Viruses. Proceedings of the National Acadacemy of Science 2004:14017-14022. 110. Poortinga AT, Bos R, Busscher HJ (1999) Measurement of charge transfer during bacterial adhesion to an indium tin oxide surface in a parallel plate flow chamber. Journal of Microbiological Methods 38:183-189. 111. Felice CJ, Valentinuzzi ME (1999) Medium and Interface Components in Impedance Microbiology. IEEE Transactions on Biomedical Engineering 46:1483-1487. 112. Sengupta S, Battigelli DA, Chang H-C (2006) A micro-scale multi-frequency reactance measurement technique to detect bacterial growth at low bioparticle concentrations. Lab on a Chip 6:682-692. 113. Li YB, Su XL (2006) Microfluidics-based optical biosensing method for rapid detection of Escherichia coli O157 : H7. Journal of Rapid Methods and Automation in Microbiology 14:96-109. 114. Varshney M, Li Y, Srinivasan B, Tung S, et al. (2006) A microfluidic filter biochip-based chemiluminescence biosensing method for detection of Escherichia coli O157 : H7. Transactions of the Asabe 49:2061-2068.
314
Sengupta, Gordon, and Chang
115. Klostranec JM, Xiang Q, Farcas GA, Lee JA, et al. (2007) Convergence of quantum dot barcodes with microfluidics and signal processing for multiplexed high-throughput infectious disease diagnostics. Nano Letters 7:28122818. 116. Koh CG, Tan W, Zhao MQ, Ricco AJ, Fan ZH (2003) Integrating polymerase chain reaction, valving, and electrophoresis in a plastic device for bacterial detection. Analytical Chemistry 75:4591-4598. 117. Lin FYH, Sabri M, Alirezaie J, Li DQ, Sherman PM (2005) Development of a nanoparticle-labeled microfluidic immunoassay for detection of pathogenic microorganisms. Clinical and Diagnostic Laboratory Immunology 12:418425. 118. Docherty FT (2004) The first SERRS multiplexing from labelled oligonucleotides in a microfluidics lab-on-a-chip. Chemical Communications 118119. 119. Ymeti A (2007) Fast, ultrasensitive virus detection using a young interferometer sensor. Nano Letters 7:394-397. 120. Hunt HC, Wilkinson JS (2008) Optofluidic integration for microanalysis. Microfluidics and Nanofluidics 4:53-79. 121. Viskari PJ (2006) Unconventional detection methods for microfluidic devices. Electrophoresis 27:1797-1810. 122. Yi CQ (2006) Optical and electrochemical detection techniques for cellbased microfluidic systems. Analytical and Bioanalytical Chemistry 384:1259-1268. 123. Gotz S (2007) Recent developments in optical detection methods for microchip separations. Analytical and Bioanalytical Chemistry 387:183-192. 124. Lazcka O, Del Campo FJ, Munoz FX (2007) Pathogen detection: A perspective of traditional methods and biosensors. Biosensors & Bioelectronics 22:1205-1217. 125. Bange A (2005) Microfluidic immunosensor systems. Biosensors & Bioelectronics 20:2488-2503. 126. Kuswandi B (2007) Optical sensing systems for microfluidic devices: A review. Analytica Chimica Acta 601:141-155. 127. Xiang Q (2006) Miniaturized immunoassay microfluidic system with electrokinetic control. Biosensors & Bioelectronics 21:2006-2009. 128. Floriano PN (2005) Membrane-based on-line optical analysis system for rapid detection of bacteria and spores. Biosensors & Bioelectronics 20:20792088. 129. Webster JR (2001) Monolithic capillary electrophoresis device with integrated fluorescence detector. Analytical Chemistry 73:1622-1626. 130. Oh KW (2005) World-to-chip microfluidic interface with built-in valves for multichamber chip-based PCR assays. Lab on a Chip 5:845-850. 131. Xiang Q (2005) Real time PCR on disposable PDMS chip with a miniaturized thermal cycler. Biomedical Microdevices 7:273-279. 132. Dittrich PS (2005) Single-molecule fluorescence detection in microfluidic channels - the Holy Grail in mu TAS? Analytical and Bioanalytical Chemistry 382:1771-1782.
Microfluidic Diagnostic Systems
315
133. Hollars CW (2006) Bio-assay based on single molecule fluorescence detection in microfluidic channels. Analytical and Bioanalytical Chemistry 385:1384-1388. 134. Meldrum DR (2002) Microscale bioanalytical systems. Science 297:11971198. 135. Higgins JA (2003) A handheld real time thermal cycler for bacterial pathogen detection. Biosensors & Bioelectronics 18:1115-1123. 136. Roederer M, DeRosa A, Gerstein R, Anderson M, et al. (1997) 8 color, 10 parameter flow cytometry to elucidate complex leucocyte heterogeneity. Cytometry 29:328-339. 137. Boeck G (2001) Current status of flow cytometry in cell and molecular biology. International Reviews in Cytology 204:239-298. 138. Huh D, Gu W, Kamotani Y, Grotberg JB, Takayama S (2005) Microfluidics for flow cytometric analysis of cells and particles. Physiological Measurement 26:R73-R98. 139. Stone HA, Stroock AD, Ajdari A (2004) Engineering flows in small devices: Microfluidics towards lab-on-a-chip. Annual Review of Fluid Mechanics 26. 140. Chang H-C (2006) Electrokinetics: a viable microfluidic platform for miniature diagnostic kits. Canadian Journal of Chemical Engineering 84:146-160. 141. Laser DJ, Santiago JG (2004) A review of micropumps. Journal of Micromechanics and Microengineering 14:R35-R64. 142. Sakamoto C, Yamaguchi M, Nasu M (2005) Rapid and simple quantification of bacterial cells using a microfluidic device. Applied and Environmental Microbiology 71:1117-1121. 143. Dittrich P, Schwill P (2003) An integrated microfluidic system for reaction, high-sensitivity detection, and sorting of fluorescent cells and particles. Analytical Chemistry 75. 144. McClain MA, Cuthbertson CT, Jacobson SC, Ramsey JM (2001) Flow cytometry of Escherichia coli on microfluidic devices. Analytical Chemistry 73:5334-5338. 145. Fu AY, Chou H-P, Spence C, Arnold FH, Quake SR (2002) An Integrated Microfabricated Cell-sorter. Analytical Chemistry 74:2451-2457. 146. Johnson P, Lebaron P, Deromedi T, Baudart J, Catala P, International Society for Applied Cytology XXII International Congress, Montpellier, France 2004. 147. Naik T, Longmire EK, Mantell SC (2003) Dynamic response of a cantilever in liquid near a solid wall. Sensors and Actuators A Physical 102:240-254. 148. Gupta A, Akin D, Bashir R (2004) Detection of bacterial cells and antibodies using surface micromachined thin silicon cantilever resonators. Journal of Vacuum Science and Technology B 22:2785-2791. 149. Zuehlke J (2007) Rapid Detection of Foodborne E. coli O157:H7 Using Piezoelectric-excited Millimeter-size Cantilever Sensors. Basic Biotechnology eJournal 2007 3:14-19. 150. Raiteri R, Grattarola M, Butt HJ, Skladal P (2001) Micromechanical cantilevel-based biosensors. Sensors and Actuators B Chemical 79.
316
Sengupta, Gordon, and Chang
151. Ilic B, Czaplewski D, Craighead HG, Neuzil P, et al. (2000) Mechanical resonant immunospecific biological detector. Applied Physics Letters 77:450453. 152. Davis ZJ, Svendsen W, Boisen A (2007) Design, fabrication and testing of a novel MEMS resonator for mass sensing applications. Microelectronics Engineering 84:1601-1605. 153. Villarroya M, Verd J, Teva J, Abadal G, et al. (2006) System on chip mass sensor based on polysilicon cantilevers arrays for multiple detection. Sensors and Actuators A Physical 132:154-164. 154. Boisen A, Keller S, Nordstrm M, Johanson A, et al. (2007) Rapid molecular detection of food and water-borne diseases. Microbiology Today 116-118. 155. Ghatnekar-Nilsson S, Forsen E, Abadal G, Verd J, et al. (2005) Resonators with integrated CMOS circuitry for mass sensing applications, fabricated by electron beam lithography. Nanotechnology 16:98-102. 156. Mulhern GT, Soane DS, Howe RT, 7th International Conference on SolidState Sensors and Actuators (Transducers ’93), Yokohama, Japan 1993, p. 296. 157. Tien NC, Jeong S, Phinney LM, Fushinobu, Bokor J (1996) Surface adhesion reduction in silicon microstructures using femtosecond laser pulses. Applied Physics Letters 68:197-199. 158. Shin S, 93rd annual meeting of the US Animal Health Association, Las Vegas, NV 1989, p. 381. 159. Berg JD, Fiksdal L (1988) Rapid detection of total and fecal coliforms in water by enzymatic hydrolysis of 4-methylumbelliferone-beta-D-galactoside. Applied and Environmental Microbiology 54:2118-2122. 160. Deland FH, Wagner HN (1969) Early detection of bacterial growth and carbon 14 labelled glucose. Radiology 92:154-155. 161. Campbell CD, Chapman SJ, Cameron CM, Davidson MS, Potts JM (2003) A Rapid Microtiter Plate Method To Measure Carbon Dioxide Evolved from Carbon Substrate Amendments so as To Determine the Physiological Profiles of Soil Microbial Communities by Using Whole Soil. Applied and Environmental Microbiology 69:3593–3599. 162. Richards JCS, Jason AC, Hobbs G, Gibson DM, Christie RH (1978) Electronic measurement of bacterial growth. Journal of Physics Series E:Scientific Instruments 11:560-568. 163. Manafi M, Kremsmaier B (2001) Comparative evaluation of different chromogenic/fluorogenic media for detecting Escherichia coli O157:H7 in food. International Journal of Food Microbiology 71:257-262. 164. Fales WH, Reilly TJ (2004) Experience with ESP Liquid Culture System and the Rapid Culture Identification of Mycobacterium avium subsp. paratuberculosis (MAP). TREK Vet Update 1:1-2. 165. Andersen KB, von Meyenburg K (1980) Are growth rates of Escherichia coli in batch cultures limited by respiration? Journal of Bacteriology 144:114-123. 166. Børsheim KY, Bratbak G, Heldal M (1990) Enumeration and biomass estimation of planktonic bacteria and viruses by transmission electron microscopy. Applied and Environmental Microbiology 56:352-356.
Microfluidic Diagnostic Systems
317
167. Gomez-Sjoberg R, Morisette DT, Bashir R (2005) Impedance Microbiologyon-a-Chip: Microfluidic Bioprocessor for Rapid Detection of Bacterial Metabolism. Journal of Microelectromechanical Systems 14:829-838. 168. Yang L, Banada PP, Liu Y-S, Bhuniya AK, Bashir R (2005) Conductivity and pH Dual Detection of Growth Profile of Healthy and Stressed Listeria monocytogenes. Biotechnology and Bioengineering 92:685-694. 169. Venugopalan S, Scherer A, Vyawahare S, CALTECH-SURF 2006. 170. Helmke BP, Minerick AR (2006) Designing a nano-interface in a microfluidic chip to probe living cells: Challenges and perspectives. Proceedings of the National Academy of Science 103:6419-6424. 171. Higuera-Guisset J, Rodriguez-Viejo J, Chacon M, Munoz FJ, et al. (2005) Calorimetry of microbial growth using a thermopile based microreactor. Thermochimica Acta 427:187-191. 172. Castellarnau M, Zine N, Bausells J, Madrid C, et al. (2007) Integrated cell positioning and cell based ISFET biosensors. Sensors and Actuators B: Chemical 120:615-620. 173. Barken KB, Haagensen JAJ, Tolker-Nielsen T (2007) Advances in nucleic acid-based diagnostics of bacterial infections. Clinica Chimica Acta 384:1-11. 174. Mothershed EA, Whitney AM (2006) Nucleic acid-based methods for the detection of bacterial pathogens: Present and future considerations for the clinical laboratory. Clinica Chimica Acta 363:206-220. 175. Ekins PP (1999) Immunoassay, DNA analysis, and other ligand binding assay techniques: From electropherograms to multiplexed, ultrasensitive microarrays on a chip. Journal of Chemical Education 76:769-780. 176. Eteshola E, Leckband D (2001) Development and characterization of an ELISA assay in PDMS microfluidic channels. Sensors and Actuators BChemical 72:129-133. 177. Paton JC, Paton AW (1998) Pathogenesis and diagnosis of shiga toxinproducing Escherichia coli infections. Clinical Microbiology Reviews 11:450479. 178. Biomerieux Industry (2007) VIDAS Technology - VIDAS E.coli O157. Accessed 21 November, 2007. http://www.biomerieux179. Ferreira JL, Eliasberg SJ, Harrison MA, Edmonds P (2001) Detection of preformed type A botulinal toxin in hash brown potatoes by using the mouse bioasssay and a modified ELISA test. Journal of Aoac International 84:14601464. 180. Proux K, Houdayer C, Humbert F, Cariolet R, et al. (2000) Development of a complete ELISA using Salmonella lipopolysaccharides of various serogronps allowing to detect all infected pigs. Veterinary Research 31:481-490. 181. Mylonakis E, Paliou M, Lally M, Flanigan TP, Rich JD (2000) Laboratory testing for infection with the human immunodeficiency virus: Established and novel approaches. American Journal of Medicine 109:568-576. 182. Lal SP, Christopherson RI, dos Remedios CG (2002) Antibody arrays: an embryonic but rapidly growing technology. Drug Discovery Today 7:S143S149.
318
Sengupta, Gordon, and Chang
183. Selvaganapathy PR, Carlen ET, Mastrangelo CH (2003) Recent progress in microfluidic devices for nucleic acid and antibody assays. In: Proceedings of the Ieee 91:954-975. 184. Stokes DL, Griffin GD, Tuan VD (2001) Detection of E-coli using a microfluidics-based antibody biochip detection system. Fresenius Journal of Analytical Chemistry 369:295-301. 185. Bai YL, Koh CG, Boreman M, Juang YJ, et al. (2006) Surface modification for enhancing antibody binding on polymer-based microfluidic device for enzyme-linked immunosorbent assay. Langmuir 22:9458-9467. 186. Gao YL, Lin FY, Hu GQ, Sherman PA, Li DQ (2005) Development of a novel electrokinetically driven microfluidic immunoassay for the detection of Helicobacter pylori. Analytica Chimica Acta 543:109-116. 187. Liu WT, Zhu L, Qin QW, Zhang Q, et al. (2005) Microfluidic device as a new platform for immunofluorescent detection of viruses. Lab on a Chip 5:1327-1330. 188. Sato K, Tokeshi M, Odake T, Kimura H, et al. (2000) Integration of an immunosorbent assay system: Analysis of secretory human immunoglobulin A on polystyrene beads in a microchip. Analytical Chemistry 72:1144-1147. 189. Haes AJ, Terray A, Collins GE (2006) Bead-assisted displacement immunoassay for staphylococcal enterotoxin B on a microchip. Analytical Chemistry 78:8412-8420. 190. Lee LJ, Yang ST, Lai SY, Bai YL, et al., Advances in Clinical Chemistry, Vol 42 2006, pp. 255-295. 191. Bai YL, Huang WC, Yang ST (2007) Enzyme-linked Immunosorbent assay of Escherichia coli O157 : H7 in surface enhanced Poly(Methyl methacrylate) microchannels. Biotechnology and Bioengineering 98:328-339. 192. Peterson DS (2005) Solid supports for micro analytical systems. Lab on a Chip 5:132-139. 193. Lee NY, Yang Y, Kim YS, Park S (2006) Microfluidic immunoassay platform using antibody-immobilized glass beads and its application for detection of Escherichia coli O157 : H7. Bulletin of the Korean Chemical Society 27:479-483. 194. Chattopadhyay S, Patra R, Ramamurthy T, Chowdhury A, et al. (2004) Multiplex PCR assay for rapid detection and genotyping of Helicobacter pylori directly from biopsy specimens. Journal of Clinical Microbiology 42:28212824. 195. Dalovisio JR, MontenegroJames S, Kemmerly SA, Genre CF, et al. (1996) Comparison of the amplified Mycobacterium tuberculosis (MTB) direct test, Amplicor MTB PCR, and IS6110-PCR for detection of MTB in respiratory specimens. Clinical Infectious Diseases 23:1099-1106. 196. Takahashi T, Nakayama T (2006) Novel technique of quantitative nested real-time PCR assay for Mycobacterium tuberculosis DNA. Journal of Clinical Microbiology 44:1029-1039. 197. Yam WC, Cheng VCC, Hui WT, Wang LN, et al. (2004) Direct detection of Mycobacterium tuberculosis in clinical specimens using single-tube bioti-
Microfluidic Diagnostic Systems
319
nylated nested polymerase chain reaction-enzyme linked immunoassay (PCRELISA). Diagnostic Microbiology and Infectious Disease 48:271-275. 198. Martin DH, Cammarata C, van der Pol B, Jones RB, et al. (2000) Multicenter evaluation of AMPLICOR and automated COBAS AMPLICOR CT/NG tests for Neisseria gonorrhoeae. Journal of Clinical Microbiology 38:3544-3549. 199. Davies HD, Miller MA, Faro S, Gregson D, et al. (2004) Multicenter study of a rapid molecular-based assay for the diagnosis of group B Streptococcus colonization in pregnant women. Clinical Infectious Diseases 39:1129-1135. 200. Bukhari Z, Weihe J, LeChevallier M (2004) Development of procedures for rapid detection of E-coli O157 : H7 from source and finished water samples. Water Science and Technology 50:233-237. 201. Lagally ET, Mathies RA (2004) Integrated genetic analysis microsystems. Journal of Physics D-Applied Physics 37:R245-R261. 202. Zhang CS, Xu JL, Ma WL, Zheng WL (2006) PCR microfluidic devices for DNA amplification. Biotechnology Advances 24:243-284. 203. Northrup MA, Ching MT, White RM, Watson RT, Tranducer '93, seventh international conference on solid state Sens Actuators, Yokohama, Japan 1993, pp. 924-926. 204. Lagally ET, Scherer JR, Blazej RG, Toriello NM, et al. (2004) Integrated portable genetic analysis microsystem for pathogen/infectious disease detection. Analytical Chemistry 76:3162-3170. 205. Morrison T, Hurley J, Garcia J, Yoder K, et al. (2006) Nanoliter high throughput quantitative PCR. Nucleic Acids Research 34. 206. Marcus JS, Anderson WF, Quake SR (2006) Parallel picoliter RT-PCR assays using microfluidics. Analytical Chemistry 78:956-958. 207. Matsubara Y, Kerman K, Kobayashi M, Yamamura S, et al. (2005) Microchamber array based DNA quantification and specific sequence detection from a single copy via PCR in nanoliter volumes. Biosensors & Bioelectronics 20:1482-1490. 208. Guttenberg Z, Muller H, Habermuller H, Geisbauer A, et al. (2005) Planar chip device for PCR and hybridization with surface acoustic wave pump. Lab on a Chip 5:308-317. 209. Neuzil P, Zhang CY, Pipper J, Oh S, Zhuo L (2006) Ultra fast miniaturized real-time PCR: 40 cycles in less than six minutes. Nucleic Acids Research 34. 210. Jones M, Alland D, Marras S, El-Hajj H, et al. (2001) Rapid and sensitive detection of Mycobacterium DNA using cepheid SmartCycler (R) and tube lysis system. Clinical Chemistry 47:1917-1918. 211. Ottesen EA, Hong JW, Quake SR, Leadbetter JR (2006) Microfluidic digital PCR enables multigene analysis of individual environmental bacteria. Science 314:1464-1467. 212. Chang YH, Lee GB, Huang FC, Chen YY, Lin JL (2006) Integrated polymerase chain reaction chips utilizing digital microfluidics. Biomedical Microdevices 8:215-225. 213. Niu ZQ, Chen WY, Shao SY, Jia XY, Zhang WP (2006) DNA amplification on a PDMS-glass hybrid microchip. Journal of Micromechanics and Microengineering 16:425-433.
320
Sengupta, Gordon, and Chang
214. Easley CJ, Karlinsey JM, Landers JP (2006) On-chip pressure injection for integration of infrared-mediated DNA amplification with electrophoretic separation. Lab on a Chip 6:601-610. 215. Gong HQ, Ramalingam N, Chen LQ, Che J, et al. (2006) Microfluidic handling of PCR solution and DNA amplification on a reaction chamber array biochip. Biomedical Microdevices 8:167-176. 216. Zhou ZM, Liu DY, Zhong RT, Dai ZP, et al. (2004) Determination of SARScoronavirus by a microfluidic chip system. Electrophoresis 25:3032-3039. 217. Kaigala GV, Huskins RJ, Preiksaitis J, Pang XL, et al. (2006) Automated screening using microfluidic chip-based PCR and product detection to assess risk of BK virus-associated nephropathy in renal transplant recipients. Electrophoresis 27:3753-3763. 218. Poulsen CR, El-Ali J, Perch-Nielsen IR, Bang DD, et al. (2005) Detection of a putative virulence cadF gene of Campylobacter jejuni obtained from different sources using a microfabricated PCR chip. Journal of Rapid Methods and Automation in Microbiology 13:111-126. 219. Wang W, Li ZX, Luo R, Lu SH, et al. (2005) Droplet-based micro oscillating-flow PCR chip. Journal of Micromechanics and Microengineering 15:1369-1377. 220. Prakash R, Kaler K (2007) An integrated genetic analysis microfluidic platform with valves and a PCR chip reusability method to avoid contamination. Microfluidics and Nanofluidics 3:177-187. 221. Burns MA, Johnson BN, Brahmasandra SN, Handique K, et al. (1998) An integrated nanoliter DNA analysis device. Science 282:484-487. 222. Pal R, Yang M, Lin R, Johnson BN, et al. (2005) An integrated microfluidic device for influenza and other genetic analyses. Lab on a Chip 5:1024-1032. 223. Kopp MU, de Mello AJ, Manz A (1998) Chemical Amplification: Continuous-Flow PCR on a Chip. Science 280:1046-1048. 224. Li SF, Fozdar DY, Ali MF, Li H, et al. (2006) A continuous-flow polymerase chain reaction microchip with regional velocity control. Journal of Microelectromechanical Systems 15:223-236. 225. Cho YK, Kim J, Lee Y, Kim YA, et al. (2006) Clinical evaluation of microscale chip-based PCR system for rapid detection of hepatitis B virus. Biosensors & Bioelectronics 21:2161-2169. 226. Shen KY, Chen XF, Guo M, Cheng J (2005) A microchip-based PCR device using flexible printed circuit technology. Sensors and Actuators B-Chemical 105:251-258. 227. Fox JT, Renter DG, Sanderson MW, Thomson DU, et al. (2007) Evaluation of culture methods to identify bovine feces with high concentrations of Escherichia coli O157. Applied and Environmental Microbiology 73:5253-5260. 228. Chang JM, Fang TJ (2007) Survival of Escherichia coli O157 : H7 and Salmonella enterica serovars Typhimurium in iceberg lettuce and the antimicrobial effect of rice vinegar against E.coli O157 : H7. Food Microbiology 24:745-751.
Microfluidic Diagnostic Systems
321
229. Vidovic S, Germida JJ, Korber DR (2007) Sensitivity of two techniques to detect Escherichia coli O157 in naturally infected bovine fecal samples. Food Microbiology 24:633-639. 230. (CDC) CfDCaP,. 2006. 231. Wu VCH, Chen SH, Lin CS (2007) Real-time detection of Escherichia coli O157 : H7 sequences using a circulating-flow system of quartz crystal microbalance. Biosensors & Bioelectronics 22:2967-2975. 232. Yang JN, Liu YJ, Rauch CB, Stevens RL, et al. (2002) High sensitivity PCR assay in plastic micro reactors. Lab on a Chip 2:179-187. 233. Xiang Q, Xu B, Fu R, Li D (2005) Real time PCR on disposable PDMS chip with a miniaturized thermal cycler. Biomedical Microdevices 7:273-279. 234. Hu GQ, Xiang Q, Fu R, Xu B, et al. (2006) Electrokinetically controlled realtime polymerase chain reaction in microchannel using Joule heating effect. Analytica Chimica Acta 557:146-151. 235. Yeung SW, Lee TMH, Cai H, Hsing IM (2006) A DNA biochip for on-thespot multiplexed pathogen identification. Nucleic Acids Research 34. 236. Belgrader P, Benett W, Hadley D, Richards J, et al. (1999) Infectious disease - PCR detection of bacteria in seven minutes. Science 284:449-450. 237. Yuen PK, Kricka LJ, Fortina P, Panaro NJ, et al. (2001) Microchip module for blood sample preparation and nucleic acid amplification reactions. Genome Research 11:405-412. 238. Waters LC, Jacobson SC, Kroutchinina N, Khandurina J, et al. (1998) Microchip device for cell lysis, multiplex PCR amplification, and electrophoretic sizing. Analytical Chemistry 70:158-162. 239. Zhang CS, Xing D (2007) Miniaturized PCR chips for nucleic acid amplification and analysis: latest advances and future trends. Nucleic Acids Research 35:4223-4237. 240. Prakash AR, Adamia S, Sieben V, Pilarski P, et al. (2006) Small volume PCR in PDMS biochips with integrated fluid control and vapour barrier. Sensors and Actuators B-Chemical 113:398-409. 241. Lagally ET, Emrich CA, Mathies RA (2001) Fully integrated PCR-capillary electrophoresis microsystem for DNA analysis. Lab on a Chip 1:102-107. 242. Wang F, Yang M, Burns MA (2008) Microfabricated valveless devices for thermal bioreactions based on diffusion-limited evaporation. Lab on a Chip 8:88-97. 243. Roper MG, Easley CJ, Landers JP (2005) Advances in polymerase chain reaction on microfluidic chips. Analytical Chemistry 77:3887-3893. 244. Min JH, Baeumner A (2003) The micro-total analytical system for the detection of bacteria/viruses. Journal of Industrial and Engineering Chemistry 9:18. 245. Auroux PA, Koc Y, deMello A, Manz A, Day PJR (2004) Miniaturised nucleic acid analysis. Lab on a Chip 4:534-546. 246. Lepp PW, Brinig MM, Ouverney CC, Palm K, et al. (2004) Methanogenic Archaea and Human Periodontal Disease. Proceedings of the National Academy of Science 101:6176-6181.
322
Sengupta, Gordon, and Chang
247. Ley RE, Backhed F, Turnbaugh P, Lozupone CA, et al. (2005) Obesity alters gut microbial ecology. Proceedings of the National Academy of Science 102:11070-11075. 248. Lakatos P, Fischer S, Lakatos L, Gal I, Papp J (2006) Current concept on the pathogenesis of inflammatory bowel disease - crosstalk between genetic and microbial factors: Pathogenic bacteria and altered bacterial sensing or changes in mucosal integrity take "toll" ? World Journal of Gastroenterology 12:18291841. 249. Petrof EO, Ciancio MJ, Chang EB (2004) Role and regulation of intestinal epithelial heat shock proteins in health and disease. Chinese Journal of Digestive Diseases 5:45-50.
Chapter 10 Microfluidic Applications in Biodefense
Stevan Jovanovich and Joanne Horn Microchip Biotechnologies Inc, 6693 Sierra Lane, Suite F, Dublin, CA 94568 Correspondence should be addressed to: Stevan Jovanovich (
[email protected])
Keywords: microfluidics, biodefense, microchip, microscale sample preparation; immunomagnetic separations, toxin detection, proteomics, microarrays, microelectrophoresis, integrated systems, detect-to-warn, detect-totreat, PCR, qPCR, lab-on-a-chip, JBAIDS, JBPDS, paramagnetic beads, spore disruption, nucleic acid purification, immunoassays, electrochemiluminescence.
Abstract There is an increasing need for compact biodefense devices that work autonomously and consume minimal reagents. These requirements can be well met by microfluidic technologies. This review first describes the needs for biodefense, which include protection of civilian populations with detect-to-warn and detect-to-treat modalities, and the needs of the military. The different microfluidic technologies applied to each step of threat detection are then discussed. The technology areas covered are microscale
324 Jovanovich and Horn
sample preparation, immunomagnetic separations, immunoassays, toxin detection, proteomics, microarrays, and microelectrophoresis. For each technology area, the potential microfluidic solutions are introduced and current relevant examples are described. For each area, the potential applications to biodefense are detailed. The present state of fully integrated microfluidic devices is reviewed. Finally, perspectives for the future are discussed.
10.1 Introduction The growing threat of bioterrorism attacks, combined with repeated outbreaks of emerging infectious diseases, underlines the importance of infrastructure improvement for the detection and diagnosis of biowarfare agents and emerging pathogens. The growing threats posed by terrorists and rogue nations⎯as evidenced by Iraq’s acknowledgement following the first Gulf war that it had loaded biological weapons, and multiple biological “incidents” worldwide⎯has raised serious concerns about bioterrorism attacks directed against the United States and other nations. Following the October 2001 anthrax attacks in the United States, biodefense has become an area of utmost national and international urgency [12]. This incident sparked the testing of tens of thousands of samples for the presence of anthrax, straining the Laboratory Response Network (LRN) system. The New York City experience after the anthrax attacks is telling [3]. The increase in incoming samples went from one every several months to about 6,000 in two weeks, requiring a coordinate growth of analytical staff and laboratories by over twenty-fold. The operational expectation had been that any surge would primarily be composed of human clinical samples; instead most of the samples were environmental. A recent study of acute care facilities in Mississippi found that the diagnostic capacity of hospitals would be overwhelmed by a weapon of mass destruction (WMD) attack [4]; similar conclusions regarding the lack of diagnostic surge capacity in alternate locales were reached in other studies [5-6]. The destructive potential of genetically engineered bioagents is huge. Toxic genes can be hidden in innocuous organisms and expressed at high levels. Expression timing and genotypic specificity could be controlled to maximize impact and potentially limit spread to a defined racial pool. The purported accomplishments of the Soviet bioweapon program [7] accentu-
Microfluidic Applications in Biodefense 325
ate how real the problem is: sophisticated state-sponsored bioweapons programs have already genetically engineered bacteria and viruses to increase their devastating impact on human populations. “Capitalizing” on post-1972 advances in biotechnology such as genetic engineering, the Soviet Union program researched and produced a range of weapons employing smallpox, anthrax, plague, and other dangerous pathogens [8]. Fortunately, terrorists and rogue states have not yet fully incorporated biological weapons into their arsenals, to our knowledge [9]. The detection of bioengineered organisms presents even greater challenges than detecting conventional pathogens, and may require multi-tiered screening, including high resolution detection of target genes and DNA sequencing. Similar to the needs for biodefense of engineered organisms, naturally emerging infectious diseases present another major threat to human health, through the natural spread of these organisms, their rapid evolution to human hosts, and the potential for bioterrorism using these agents. The recent outbreaks of severe acute respiratory syndrome (SARS) and bovine spongiform encephalopathy (BSE), commonly known as mad-cow disease, have raised global concerns on the need for rapid identification of causative agents and infected individuals before the virus spreads beyond control. Hantavirus pulmonary syndrome and West Nile virus are examples of additional infectious diseases now emerging in the US. Rapid molecular diagnostic methods and monitoring platforms that can adaptively be configured for newly emerging infectious diseases or newly engineered bioagents will be essential for combating these diseases and for biodefense applications. Past incidents and the dangers of future bioterrorism attacks highlight the critical need for improved field- and laboratory-based systems to detect, identify, and subtype bioagents [10-11]. As will be seen, the state-of-art of biodefense systems today is operational but rudimentary: all US mail is screened at sorting centers. There is a strong demand from the U.S. government for next-generation systems for civilian and military applications. Biodefense monitoring equipment has even more stringent requirements: the size of the current equipment, and the acquisition and operating costs place severe constraints on widespread implementation and deployment. New analytical systems are needed that are scalable, more automated, and capable of rapid deployment in response to surging needs or field operations. The equipment needs to be smaller, use less reagents, be simpler to use, more integrated, and automated—all attibutes of microfluidic systems, making microfluidics an ideal platform to fulfill biodefense needs.
326 Jovanovich and Horn
This review summarizes approaches to detecting and characterizing biological threat agents for both civilian and military biodefense; describes biodefense programs in place and under development; and delineates some of the approaches where microfluidics is presently being applied to monitor, detect, and characterize biothreat agents. Some of the primary needs are outlined, and the challenges to design and build fully integrated microfluidic systems are described. The majority of the review surveys microfluidic technologies that might or could be used in future biodefense systems. Sections cover microscale sample preparation methods; immunomagnetic separations and immunoassays; proteomics; polymerase chain reaction (PCR), quantitative PCR (qPCR) and other nucleic acid amplification methods; DNA microarrays, microelectrophoresis, and finally integrated Lab-on-a-Chip systems.
10.2 Biodefense Monitoring There are two basic biodefense detection approaches: detect-to-warn and detect-to-treat. Detect-to-warn systems aim to identify biothreats rapidly enough to provide sufficient warning to prevent exposure by the threat. Detect-to-treat systems aim to identify the causative agent for diagnostic purposes and thereby to direct healthcare workers to the most effective treatment as quickly as possible. For all systems, low false alarm rates (FAR) and affordable acquisition and operating costs are essential for widespread adoption.
10.2.1 Civilian biodefense Civilian biodefense is based upon surveillance to detect biothreat agents, response networks to warn and direct the treatment of the affected population [12], and the development of countermeasures [13]. Bioterrorism incidents, releases of a bioagent in a form that can harm individuals or larger populations, can range from mailing of toxin or bacteria, to release of aerosols in high profile events, to attacks on the food supply. BioShield, a countermeasures program [14], is a 10-year, $5.6 billion U.S. program for the advanced development and purchase of medical countermeasures. Acquisition programs have been announced to counter Bacillus anthracis (anthrax), Variola virus (smallpox), botulinum toxins, and radiological/nuclear agents.
Microfluidic Applications in Biodefense 327
The primary programs that have been implemented for civilian biodefense detection include screening of all postal mail at sorting centers, the BioWatch program, selected localized screening in subways and other undisclosed locations, and the LRN. The currently deployed systems principally use full volume or meso-scale fluidics. Despite the need for more advanced detect-to-warn biodefense detection for the general public, these systems have largely remained undeveloped in large part due to the complexity of integrating the complete process. The largest monitoring program is BioWatch [15-16], a joint effort by the Department of Homeland Security (DHS), the Centers for Disease Control and Prevention (CDC), and the Environmental Protection Agency (EPA). BioWatch is a ‘detect-to-treat’ program and monitors the air in at least 31 cities (and as many as 120 cities) [17] for significant release of bioagents. DHS does not report cities monitored by the BioWatch program or the assays used. In 2004 and 2005, DHS funding for the Biowatch was $26.8 million for the EPA and $28.5 million for the CDC [18]. The EPA is responsible for continual air sampling by aerosol collectors that trap airborne particles onto filters. The filters are collected for analysis every 24 hours. The CDC is responsible for the analysis of the filters at state and local public health laboratories, and developing new protocols in coordination with both the Department of Energy (DOE) national laboratories and the EPA. The assays are generally acknowledged to be PCR amplified detection of specific targets using classified primer sequences. The list of target agents is similarly classified but thought to include at least anthrax, smallpox, plague, and tularemia. The BioWatch program has now processed over 2 million samples without a false positive [19]—an impressive accomplishment. However, the daily sampling frequency and the amount of coverage of the Biowatch program still leaves the civilian populace vulnerable. DHS has been funding the Biological Autonomous Networked Detectors (BAND) program as a next-generation system to alleviate some of the deficiencies of the BioWatch program. The original goals of the BAND program were continuous air monitoring with sample analysis every three hours. The detection limits were 100 organisms or 10 ng of toxin per 17,000 liters of air processed, with a very low FAR. The instrument was planned to have an acquisition cost of $25,000, an annual operating cost of $15,000, and run autonomously for 30 days. Additional requirements were for dimensions of 2 ft3 and the ability to detect up to 20 organisms and tox-
328 Jovanovich and Horn
ins. The BAND participants have almost uniformly taken a microfluidics approach. Their efforts are described in more technical detail in a later section of this chapter. In addition to detecting bioagents in the primary environment, a major effort has gone into upgrading the response of the medical community to detect unannounced attacks [20], since the expectation is that in many scenarios the first alert will be in the form of patients presenting at doctors’ offices or hospitals [21]. Biodefense systems are also required to monitor food and water sources [22-23], suspect powders, the exposure of first responders, and test for decontamination after treatment of personnel, equipment, and key environments. 10.2.2 Military biodefense The U.S. military biodefense programs aim to detect and identify biological warfare agents that an enemy might use to degrade forces, contaminate bases, and spread confusion throughout command and control systems. Various defense programs are delivering technologies that are beginning to counter these vulnerabilities. For example, the Portal Shield program is designed for facilities protection; the Joint Biological Agent Identifiation and Diagnostic System (JBAIDS) is designed for both detection and diagnostics of environmental and clinical samples [24], and the Joint Biological Point Detection System (JBPDS) is designed for detect-to-warn capabilities in the field. Portal Shield is an array-able sensor system developed to provide early warning of biological attacks for high-value, fixed-site assets, such as air bases and port facilities. Portal Shield is designed to detect and identify threats simultaneously within 25 min. It is programable to survey continuously as well as perform random or directed sampling. Portal Shield was deployed in the Persian Gulf region in February 1998 during Operation Desert Thunder, and the current instrumentation is about two thirds the size of a typical office desk. It's fully modularized, selfcontained, and can detect eight different agents. As many as 18 sensor units may be arrayed in a given area and are able to communicate with each other, so there is no reliance on just one of them sounding an alarm. Using an array system, the false positive rate diminishes towards zero.
Microfluidic Applications in Biodefense 329 Table 10.1 JBAIDS targeted agents [24] Agent or Disease Block I Anthrax Brucellosis Ebola-Marburg Plague Q fever Salmonellosis Smallpox Tularemia Typhus fever Block II Botulinum Ricin Staphylococcus enterotoxins (e.g., staphylococcal enterotoxin B)
Organism Bacillus anthracis Brucellae Filoviridae Yersinia pestis Coxiella burnetii Salmonellae Orthropox viruses Franciscella tularensis Rickettsiaei Clostridium botulinum Ricinus communis Staphylococcus aureus
JBAIDS is the DoD’s first common platform for identification and diagnostic confirmation of biological agent exposure or infection. JBAIDS Block I is currently operational as a real-time PCR instrument with FDA approved assays. The current JBAIDS Block I system (and JBTDS systems) utilizes manual sample preparation (disrupting the cell or spore and extracting the nucleic acid or protein of interest) which results in: 1) complex operator procedures that may result in human error, 2) increased operational costs, and 3) support and logistical requirements that preclude remote operations. The intent of JBAIDS Block III, Next Generation Diagnostics (NGD), is to establish a new system incorporating the capabilities of Block I and Block II capabilities (Table 10.1) and adding immunoassay capabilities and the ability to identify up to 50 agents including toxins in 15 minutes using automated, miniaturized sample preparation integrated with analysis for nucleic acids and proteins, in a hand held or smaller format. JBPDS is planned to detect, identify, and warn against the presence of up to 24 biowarfare agents at discrete points within a given field environment. JBPDS is being designed with a sampler, trigger detection, and identification technologies that allow it to rapidly and automatically detect and identify biological threat agents. JBPDS was committed to initial limited production and procurement during FY2006, with planned full production in 2007. Future JBPDS improvements will reduce size, weight,
330 Jovanovich and Horn
power consumption, and reagent use while increasing the number of agents recognized, sensitivity, and system reliability. A number of the requirements of the above programs may be met through the implementation of microfluidics. For example, reduced consummable use, weight and size are directly tied to the drastic reduction of reaction volume employed in microfluidic systems and the use of miniaturized microfluidic components. Increasing the number of recognized agents can also be accomodated without high order reaction multiplexing by using multi-channel microfluidic devices with single-plex reactions carried out in parallel. In theory, with microfabricated microchips adding more channels is simple a ‘cut and paste’ exercise once the issues of connections and integation are solved. There are considerable issues in integrating all processing and analysis steps in a ‘hands-free’ device. However, microfluidics requires a ‘hands-free’ implementation once samples are loaded since typically there is nothing the user can do to intervene.
10.3 Current Biodefense Detection and Identification Methods The detection of biothreat agents today can be segmented into laboratory detection and field detection methods, and by the type of sample matrix processed (reviewed in [11,25-28]). The detection must be sensitive, specific for the test organism(s), and may require substantial up-front sample preparation before the read-out assay can be performed. There are a number of commercial tests available today to detect biothreats using nucleic acid, immunological, and biochemical methods. The appropriate test may be determined by the level of information required (i.e., phenotypic or genotypic), timeframe, and the consequences. In general, nucleic acid tests are more sensitive, but require a higher order of skilled operators and more sophisticated equipment than immunological tests, and cannot detect toxins. Confounding the problem is the need to assess samples for the presence of many possible biothreat agents; the detection of genetically engineered organisms potentially designed to evade standard detection methods; and the vast numbers of people crossing international borders (600 million international travelers per year [27]).
Microfluidic Applications in Biodefense 331
Sample matrix and sample preparation methods are key variables in biodefense detection. Targets may be contained in air, food, water, bodily fluids, powders, swabs, swipes, cloth, or filters, among many other possibilities. Significant reductions in sample volume are required and sample preparation methods must remove contaminants such as metal ions, heme, humic acids, and other compounds, which inhibit PCR or other assays. To enable input of samples into microfluidic systems, the target analyte needs to be highly concentrated without concentrating inhibitors. Sample preparation issues are further detailed in section 10.5. 10.3.1 Laboratory detection Traditional identification of pathogens is often tedious and prolonged, involving batteries of tests that often take days or even longer to confirm. Standard clinical laboratory identification of Bacillus anthracis serves as an example [12]. Bacillus anthracis testing by a LRN Level A laboratory begins with growth of the organism, a Gram stain, capsule observation and routine culturing on sheep blood agar. This is followed by observation of colony morphology, motility, sporulation, and hemolysis. The presumptive identification can take up to 96 hours with additional days for confirmation. Confirmatory testing at Level B labs (State and Federal laboratories) consists of phage lysis and immuno-fluorescence assays of a cell wall and capsule. Level C laboratories determine antimicrobial susceptibility and apply more advanced technology including PCR, qPCR, and time-resolved fluorescence measurements. Finally, at Level D, labs (CDC), in-depth molecular characterization is performed using multiple-locus VNTR (variable-number tandem repeat) analysis (MLVA), 16S rDNA ribotyping and other methods, including sequencing to provide subtyping information for identification [29]. Full characterization can include complete genomic sequencing to identify the exact strain variant for epidemiological and forensic analysis. Similarly the Armed Forces Institute of Pathology uses DNA extraction, DNA quantification, qPCR of unique genetic targets, 16S rRNA gene sequencing, amplified fragment length polymorphism polymerase chain reaction (AFLP-PCR), and repetitive element polymerase chain reaction DNA fingerprinting to characterize strains [30]. The results are compared to extensive databases that have been assembled [31-32].
332 Jovanovich and Horn
10.3.2 Field detection For more rapid detection in the field, current detection methods use immunoassays, PCR, and other molecular typing methods to provide information on suspected biothreat samples. Direct fluorescence assay with monoclonal antibodies to cell wall and capsule components have also worked well for B. anthracis [33]. Real-time PCR or qPCR methods are frequently performed [29] and can yield rapid identification in the field, using the LightCycler (Roche), GeneXpert (Cepheid), JBAIDS (Idaho Technology) or other systems. DNA sequencing of 16S ribosomal sequences, plasmids, or variable regions can yield the highest resolution identification, but is relatively slow and has a low throughput compared to other methods; sequencing has not yet been adapted to field applications. One of the most advanced field detection concepts is qPCR packaged in portable devices. Researchers at Lawrence Livermore National Laboratories (LLNL) demonstrated real-time detection of PCR products in a miniaturized silicon reactor with thin film heaters and integrated fluorescence detection [34]. This work was extended at LLNL to a 10-channel advanced nucleic acid analyzer and a portable version was devised [35]. Cepheid developed commercial versions of this device with integrated (multimicroliter) sample processing for qPCR analysis, now incorporated into the Northrop Grumman Biohazard Detection System (BDS), used by the Postal System for monitoring mail facilities. The BDS incorporates upstream air sampling with the GeneXpert (Cepheid), which prepares DNA from a fluidized aerosol sample and then performs qPCR. The system is fully integrated, automated, and reports data to a central point. The BioWatch program grew in part from the Biological Aerosol Sentry and Information System (BASIS) project developed at the LLNL and Los Alamos National Laboratory. BASIS used filters to collect aerosolized particulate samples at large events such as the Olympics. Filters were analyzed by PCR in separate laboratory facilities using largely manual protocols [36]. LLNL also developed the Autonomous Pathogen Detection System (APDS) [37] as a stand-alone, autonomous aerosol detection device. APDS, a podium-sized system, monitors air for all three biological threat agent types (bacteria, viruses, and toxins) by continuously performing aerosol collection, sample preparation, and multiplexed biological tests. The APDS first employs fluorescent bead-based immunoassays to detect more than ten agents. If a positive signal is detected, a second tier confirmation using qPCR is enabled. APDS systems have been field tested at major transportation centers and at special events [32].
Microfluidic Applications in Biodefense 333
A number of companies (e.g., Tetracore, Alexeter Technologies) have developed and market lateral flow immunoassays for field detection of a number of relevant biothreat agents. These are strips of porous materials to which sample is added to one end, and during migration to the other end, encounters a region containing antibodies specific to the target along with a chromophore. While most immunoassays are not particularly sensitive, they are the normally used field screening method and are also used in situations where large numbers of negative samples are anticipated.
10.4 Microfluidic Challenges for Advanced Biodefense Detection and Identification Methods Rapid detection, identification, and subtyping analysis of pathogenic organisms and toxins are critical needs for biodefense and for the management of emerging infectious diseases. Autonomous systems that can detect and provide initial identification of bioagents are required for field monitoring and to provide any reasonable degree of protection to civilians. Laboratories require automated systems that can rapidly genotype microorganisms from human samples, environmental samples, or food and differentiate at the strain level and better direct treatment. All of these newly developed systems must detect nucleic acids and/or toxins in varying amounts, formats, and in many different matrices. They will need to be completely automated or simple to use; incorporate advanced technologies including sample preparation starting from primary samples (aerosols, blood, etc.), molecular detection, automation, microfluidics, and bioinformatics; reduce reagent consumption and space requirements; and provide cost and performance advantages compared to present systems. Analytical techniques such as PCR, VNTR, MLVA, AFLP, and single molecule detection are well suited to analysis on microfluidic systems. In addition, the systems should be capable of accommodating new assays as they become available. The burgeoning field of microfluidics can offer remedies that fulfill many of these needs and is thus becoming an ever increasingly desired component of next-generation systems. Microfluidics can provide the fundamental platform technology that reduces the footprint, minimizes reagent consumption, and fully automates monitoring and analytical equipment for operation in the field, monitoring of cities, and detection in the
334 Jovanovich and Horn
clinic. However, microfluidics faces many challenges before it is ready for wide-spread deployment. The first challenge is interfacing microfluidics with the full scale samples input at the front end. ‘Real world’ samples may be measured in milliliter volumes while microfluidics typically manipulates nanoliter volumes: therefore the sample must be concentrated or many orders-of-magnitude of detection sensitivity will be sacrificed at the front end of the system. In addition, for biodefense and many other applications, a potentially dilute target must be detected from what can be very large volumes. The BAND program specification for civilian protection is 100 organisms in 17,000 liters of air. To work in a microfluidic system, the ‘real world’ input sample must therefore be reduced in volume by orders-of-magnitude while still achieving the necessary signal-to-noise to maintain sensitivity. This means taking 100s of microliters of liquids or thousands of liters of air and concentrating the input sample into nanoliters before further microfluidic processing and reactions can take place. This can be achieved by chemically, biochemically, or physically concentrating the sample. Paramagnetic beads provide one elegant solution to both the ‘macro-tomicro’ interface and specificity. The beads, typically about several microns, can specifically or non-specifically capture nucleic acids, cells, viruses, or toxins, from large volumes of solution and move samples from the full volume world of milliliters into a hundreds of nanoliters. When a capture chemistry such as immunomagnetic separations or hydridization is performed with the beads, they can extract the desired target from high backgrounds and clutter. Paramagnetic beads simplify sample handling in the microfluidic world by minimizing the positioning demands on the fluidic system since magnetics can be used to recapture beads at any location and potentially eliminate diffusional losses. Once the sample has been introduced into the submicroliter realm, the next challenge is integrating the workflow steps in a microfluidic device. As we will see, microfluidics has been applied to the individual processes and proof-of-concept publications on almost any conceivable individual step are a proof of the potential of microfluidic approaches. However, while numerous microfluidic components are well developed in academic or research settings, a key challenge for microfluidics is to either (1) fully integrate all processes to build complete ‘sample-to-answer’ microfluidic systems, or (2) seamlessly interface microfluidic components with each other and with ‘full scale’ components into a complete system. The later requires interfacing components (or modules)—such as upstream samples from
Microfluidic Applications in Biodefense 335
aerosol collectors, swabs, and blood—with sample preparation components and downstream analytical devices, such as real-time PCR or mass spectroscopy. Ideally best of breed modules could be interfaced once microfluidic interfaces and connections are standardized. In any case, full process integration within a microfluidic system or module must be accomplished. The integration requires the coupling of different reactions, which may have multiple steps of sample purification, reagent addition, mixing, separation, and detection. Microvalves and micropumps are invaluable to isolate processes or reactions as individual steps and to move fluids to integrate different steps into a workflow. The fourth challenge is designing manufacturable biodefense microfluidic systems. The system should be lower in costs to build and operate, preferably by an order-of-magnitue. The ‘valley of death’ from proof of concept to product must be crossed and scalable manufacturing capabilities must be implemented with only low volumes as early adopters provide the initial orders. The commercial challenge of crossing the valley of death without prior governmental commitment is substantial. The focus of the remainder of this review is on the microfluidic technologies that can provide advanced rapid detection and identification of bioagents. We review microfluidic technologies that may be used with an emphasis on critical battlefield needs and civilian biodefense. In each section, the technology is briefly introduced, how microfluidics is being applied to advance the state of the art is reviewed, and how this is being applied to biodefense is described. In the final sections, fully integrated microfluidics systems are assessed and some of the devices under development for biodefense are presented. Other chapters in this book review the general state of microfluidics and many of the different formats having components that might be applicable to biodefense.
10.5 Microscale Sample Preparation Methods Starting from the sample, microfluidics can be applied to lyse organisms, concentrate, pre-separate, and purify components for further processing. State-of-the-art rapid, automated, miniaturized, modular universal sample preparation systems are required to prepare nucleic acids and proteins from biological samples in order to detect and identify high priority bioagents.
336 Jovanovich and Horn
This section describes some of the technologies that can be drawn upon to create a front-end that interacts with ‘real world, full volume’ samples. 10.5.1 Spore disruption Analysis of intracellular nucleic acids or proteins requires that cells be disrupted by physical (sonication, heating, or bead beating) or chemical means. The most challenging biothreat organisms to disrupt are the Gram positive bacterial spores, Bacillus anthracis and Clostridium botulinum. Sonication and bead beating are the most common ways to disrupt spores today. Microfluidic-based cell disruption using sonication has been reported. Belgrader and coworkers from LLNL reported in 1999 the development of a mini-sonication device that disrupted spores in 30 s; when coupled with a mini-chip PCR instrument, the complete analysis took 30 min [38]. This work was further extended at Cepheid where spores were lysed by a sonication horn (in conjunction with glass beads) through a flexible interface using a pressurized microfluidics cartridge [39] then the sonicated lysate was PCR amplified after reagent addition in a disposable cartridge [40]. Marentis et al. developed a piezoelectric microfluidic mini-sonicator and determined that it could lyse eukaryotic cells and spores with an efficiency of 50% lysis of B. subtilis spores in 30 s [41]. Laser induced disruption in a polydimethylsiloxane (PDMS) microchip has been reported for B. atrophaeus spores using a laser absorbing matrix at fluencies below 20 mJ/cm2 and without matrix above that level [42]. Small laser diodes and carboxyl-terminated magnetic beads have been shown to disrupt E. coli, Gram-positive vegetative bacteria, and hepatitis B virus mixed with human serum in a microchip with real time detection [43]. 10.5.2 Pre-separations Following lysis, the cellular material often requires separation into component fractions. Beads, gels, and membranes can also be incorporated to perform pre-separations to remove inhibitors or concentrate samples. They can provide high mass transfer rates and be made from polymer, silica, or other substrates for microchip liquid chromatography and electrochromatography applications. Agilent has developed a commercial polymer microfluidic chip for HPLC separation using ablation of polyimide to form channels [44]. A microfluidic technique which is increasingly applied is
Microfluidic Applications in Biodefense 337
the use of monoliths as a stationary phase (reviewed in [45-47]). The monoliths can be made with a variety of surface chemistries, pore sizes, and functionalized coatings. Dielectrophoretic separations can also be performed on microdevices [48]; in this regard, the Nanogen digital array will be discussed in the Microarray Section. Microchips have used dielectrophoretic separations to separate cells and nuclei with on-chip micropumps [49], and to separate erythrocytes [50]. Dielectrophoretic separations can be combined with microbeads to increase the local concentration for enhanced bead binding [51]; with microfluidic flow cytometry to sort cells [52]; with immunocapture to assay components [53]; and with ultrasonic standing waves [54]. In the future, dielectrophoresis may be applied to separate cells from debris and environmental contaminants for biodefense purifications. 10.5.3 Nucleic acid purifications Biodefense samples are derived from a wide variety of substrates and matrices. The matrix may contain complex mixtures including inhibitory compounds (e.g., mold, hemes, indigo, humic and fulvic acids, chelating agents, DNases, RNAses, and proteases) that interfere with DNA amplification, the gold standard for bioagent identification. The DoD Critical Reagents Program now provides standardized test kits composed of commonly encountered inhibitory compounds to aid biodefense development and testing. A number of approaches have been taken to purify nucleic acids before analysis at a full volume. Early work showed that dilution of the sample can relieve inhibition contained in soil extracts for PCR amplification reactions [55]. However, dilution may not be an option if target concentrations are low. Low-melting-temperature agarose has been used to extract DNA from soil samples [56]. Solid phase extraction that adsorb analytes onto columns, beads, and surfaces and spun separation gels in column format are commonly used to purify DNA before analysis. Multistep purifications such as organic extracts combined with Sephadex columns have also been developed. While these methods are effective, they were best suited for research laboratory environments due to their reliance on supplementary equipment, trained personnel, and time-consuming procedures [57]. Most are not amenable to a microfluidic format.
338 Jovanovich and Horn
Microfluidics and microchips are now being applied to miniaturize DNA extractions and concentrate nucleic acids. The surfaces of capillaries [58], beads [59], and microchips have been used with chaotrophic agents to concentrate, and purify DNA in microsystems. Microchips with silicon pillars [60-61], or plastic microchips with silicon dioxide coatings [62] can purify DNA using chaotrophic agents, ethanol wash, and elution in water. Cepheid reported nucleic acid extraction efficiencies of about 50% and concentrated samples about ten-fold for PCR [61]. Silicon microfluidic channels have been modified with amino silanes and DNA selectively eluted with alkaline rinses [63]. Immobilized beads in microchannels increased the extraction efficiency from serum 88-fold compared with free beads [64]. An integrated microfluidic device was developed to pre-treat whole blood samples using a micro-filter, micro-mixer, micro-pillar array, micro-weir and porous matrix [65]. Sol gels in capillary chromatography have been reviewed [66], as have monoliths for preconcentration and sample extraction [67]. The Landers group reported silica bead purification of DNA with chaotrophic agents and sol gel immobilization [68] on microchips, and demonstrated purification of B. anthracis DNA [69]. They have demonstrated a silica-based monolithic column in a fused-silica capillary [70] and on a glass microchip [71] with extraction of model DNA from complex samples with efficiencies of 70%. For whole blood and other mixtures they combined a C18 reverse phase column, to remove proteins and other compounds, with a monolithic column to create a dual phase sample preparation microchip [72-73]. The authors of this chapter have been developing a device, BeadStorm™, for automated magnetic bead purification technology on microfluidic handling on microchips (Fig. 10.1) at Microchip Biotechnologies Inc. The plastic sample processing cube, about 1 in³, with a 800 uL processing chamber is integrated with pneumatically actuated microvalves on a glass microchip to direct pressure-driven flows consisting of fluids, beads, and samples among reagent and reaction reservoirs. Microvalves replace both conventional valves and tubing between reservoirs providing a leak-free, self-contained fluid transport. The BeadStorm module manipulates input liquid and swabs of biological samples by bead purifying the samples, fully preparing them for downstream analysis such as PCR and immunoassays. Immunomagnetic separations have also been performed in this format. The BeadStorm device has been successfully used to automate DNA extraction from buccal swabs for STR amplification. Liquid blood samples have also been successfully prepared by the BeadStorm module
Microfluidic Applications in Biodefense 339
with automated preparation of DNA from liquid blood sample in less than 5 minutes.
Fig. 10.1 BeadStorm device (top panel) and plastic disposable ‘cube’ containing a glass chip with microvalves on the bottom to direct flows (bottom panel).
10.6 Immunomagnetic Separations and Immunoassays Immunological techniques are widely used for rapid purification and detection of cells, viruses, and proteins and are the most widely used diagnostic method in clinical medicine. The primary antibody is typically attached to a solid surface such a microtiter plate or a bead. The secondary antibody is added to generate a ‘sandwich’ assay that can provide a highly specific and rapid readout. The antibodies can be polyclonal, with wide variation of specificity from batch-to-batch, or monoclonal, which can be produced repeatedly with identical avidity to the corresponding antigen. Sandwich assays are well developed for use in laboratories, clinics, and field applications, and are the most common type of assays for toxin testing. There are many immunoassays that are being applied to biodefense and microfluidic immunoassays will be increasingly important in the future.
340 Jovanovich and Horn
10.6.1 Immunomagnetic separations Immunomagnetic separation (IMS) is a powerful technology that allows targets to be captured and concentrated in a single step using a mechanically simplified format that employs paramagnetic beads and a magnetic field (reviewed in [74-76]). IMS is used to capture, concentrate, and then purify specific target antigens, proteins, toxins, nucleic acids, cells, and spores in a single step. IMS works by binding a specific affinity reagent, typically an antibody, to paramagnetic beads, which are only magnetic in the presence of an external magnetic field. The beads can be added to complex samples such as aerosols, liquids, bodily fluids, or food. After binding of the target to the affinity reagent (which itself is bound to the paramagnetic bead) the bead is captured by application of a magnetic field. Unbound or loosely bound material is removed by washing purifing the target from other, unwanted materials in the original sample. A similar approach can purify nucleic acids using a complementary nucleic acid strand attached to a bead. Because beads are small and bind high levels of target, when the beads are concentrated by magnetic force, they form bead beds measured in the hundreds of nanoliters (or as low as a single bead), thus concentrating the target at the same time it is purified. The purified and concentrated targets can be conveniently transported, denatured, lysed or analyzed on-bead, or eluted off-bead for further sample preparation and analysis. Immunomagnetic separations are commonly used as an upstream purification step before qPCR, electrochemiluminescence, and magnetic force discrimination. As mentioned, paramagnetic beads provide an excellent solution to the macroscale-to-microscale interface: beads are an almost ideal vehicle to purify samples at the macroscale from large volumes and concentrate the specific biomolecules or targets to the nanoscale for introduction into microfluidics devices. Immunomagnetic separations are used widely for the detection of microorganisms in food and agriculture. Typically, immunomagnetic beads coated with the appropriate antibody are added to material that had been homogenized in a stomacher. The pathogenic strain Escherichia coli O157:H7 can be detected directly in ground beef with greater sensitivity than traditional plate enumeration methods by using IMS before plating [77]. IMS has also been coupled with qPCR [78], fluorescence microscopy, and solid-phase laser scanning cytometry [79]. Key parameters affecting detection sensitivity were shown to include the amount of nanoparticles per assay, immunoreaction incubation time, concentration of the target organism, matrix
Microfluidic Applications in Biodefense 341
background, and interferents. Generally, interference by non-target microorganisms is less than 1% [80]. IMS can also be combined with PCR to effectively increase the specificity of the overall process by combining the specificity of antibody-antigen recognition and the sensitivity of PCR. Immuno-quantitative PCR for S. aureus enterotoxin B (SEB) detection was found to be 1,000 times more sensitive than enzyme linked immunosorbent assays (ELISA), with little cross-reactivity [81]. Immunomagnetic separations have been adapted to microfluidic devices. In one application, electromagnetic gradients were generated using Cu micro-coil arrays embedded in a silicon substrate with magnetic pillars composed of NiCoP alloy with an integrated sensing coil to produce tunable, localized magnetic forces that were able to trap up to 80% of applied particles [82]. A miniature, integrated microfluidic device to separate magnetic particles from laminar flows was developed by the Whitesides group and demonstrated to be effective in separating live E. coli on magnetic beads [83]. Clinical applications development includes capture of targeted T cells from blood in bead beds contained in microfluidic channels. These studies resulted in 20-37% T cell capture, but only allowed flow rates of 3 uL/min [84]. Rates of hundreds of microliters per minute are required for most applications that can be achieved by increasing flow rates or by upstream deployment of IMS or other technologies that specifically capture and concentrate targets. 10.6.2 Immunoassays In addition to performing IMS purifications, extensive work has been done on developing and integrating complete immunoassays in microfluidic devices. Immunoassays include immunochromatographic lateral flow devices, ELISA, IMS-electrochemiluminescence, time-resolved fluorescence, and magnetic force assays. In addition to widespread clinical applications, immunoassays are routinely used to detect biosecurity threats [85-87]. Future immunoassays will continue to exploit advances in antibody production and screening, miniaturization, integration, and multiplexing [88]. 10.6.2.1 Lateral flow
Lateral flow assays are very simple and compatible with portable applications. A drop of test solution in buffer is added to a pad containing antibodies coupled to colloidal gold or other labels. The antibody and antigen,
342 Jovanovich and Horn
if present, bind and wick down the pad laterally and intercept detection lines that contain capture antibody-gold complexes. The gold aggregates due to the bivalency of the antibodies and produces a line that is visible by eye. Home pregnancy tests are probably the best-known lateral flow devices. Lateral flow devices are simple to use, require little training or equipment, and have been devised to detect biothreat agents. Simple dipstick immunoassays for E. coli O157:H7 can detect 1 cell per g in ground beef, after outgrowth of the sample [89]. A comparison of lateral flow assays in a handheld device with ELISAs and PCR found the sensitivity of the lateral flow assays were approximately one-hundredth of ELISAs, which were in turn one-tenth the sensitivity of PCR assays [90]. Lateral flow devices are being improved with microfluidic technologies. Monolithic beds have been combined with electrophoretic separations to produce a fast (<10 min), sensitive assay for saliva analysis [91]. An ELISA detected staphylococcal enterotoxin B (SEB) in a handheld assay for food at about 50 pg/g of matrix using lateral flow [92]. An integrated microfluidic device with sample preparation (filtration and mixing) has been described to detect botulinum neurotoxin directly from whole blood [93]. 10.6.2.2 ELISA
ELISA is the most commonly used form of immunoassay. ELISAs use an antibody bound to a solid phase support, such as a microtiter plate, to capture analytes from liquids. After washing away unbound material, a secondary antibody with a label or coupled to an enzyme (e.g., horseradish peroxidase) is used to produce a visual or fluorescent readout. ELISAs are relatively inexpensive, scalable to 96- and 384-well technologies, and can be sensitive and specific, depending on the antibody pairs. Early work with polyclonal antibodies demonstrated detection of the enterohemorrhagic E. coli O157:H7 in food at about 1 cell/g sensitivity [94]. Anti-S. aureus enterotoxin B (SEB) antibody has been immobilized on carboxylated polystyrene microparticles and a competitive assay between FITC-labeled SEB was developed to detect 0.125 ng/mL of SEB in drinking water and 0.5 ng/mL in whole milk [95]. SEB has also been assayed by direct labeling of secondary antibody with detection limits of 100 pg/well [96-97]. Immunoassays are being adapted to microfluidic devices. Microfluidic molded silicone integrated devices have successfully detected botulinum
Microfluidic Applications in Biodefense 343
neurotoxin serotype A with results equivalent to full volume assays [98]. Microfluidic immunoassays in plastic devices offer the affordability of plastics, the availability of diverse microfabrication methods, and many well-developed polymer surface modifications [99]. A heterogeneous immunoassay with antigens immobilized on PDMS-coated glass microchips with electrokinetic-control for multiple analyte detection had detection limits for E. coli O157:H7 of 3 ug/mL in an automated prototype [100]. A poly(methyl methacrylate) (PMMA) microfluidic immunoassay device was modified with poly(ethyleneimine) (PEI), an amine-bearing polymer, to increase antibody binding ten-fold [101]; the authors believe this is due to the spacer effect as well as the addition of amine groups. Due to the smaller dimensions, the microchip reactions were ten-fold faster than 96 well plates and had a dynamic range of 5 to 1000 ng/mL. An early study using glass capillary tubes as a solid support to assay E. coli O157:H7 employed a competitive-based immunoassay and achieved a detectable limit of 1 cfu per 10 g of ground beef [102]. Capillaries have also been employed to separate complex matrices and detect a model antigen at 10 pM with immunoaffinity chromatography using dual syringe pumps, a silica bead packed bed, and laser-induced fluorescence [103]. Beads have been used on microchips to separate binding steps from the secondary detection to reduce background [104] and to mix immunoassays on microchips [105]. Microfluidic immunoassays are being adapted for biodefense needs. Liu and coworkers [106] developed multi-stage integrated microfluidics for immunoassays utilizing electrochemical detection. Micropumps and circuits were integrated to perform parallel immunoassays for model organisms with enzyme-generated signal detected by active CMOS circuitry with resulting sensitivity in the fM range [106]. A Multi-Analyte Array Biosensor (MAAB) was developed at the Naval Research Laboratory (NRL) to detect multiple target agents in complex samples using a novel fluidics cube module to control the flow of solutions over six different immunoarray sensors in a small portable device with an evanescent wave detector [107]. The MAAB could rapidly detect three toxins: ricin, staphylococcal enterotoxin B, and cholera toxin [108], and S. typhimurium [109]. Liquid array-based immunoassays with multiplexed detection have also been developed and tested for model organisms [110]. Yang and coworkers at Nanogen exploited the unique characteristics of their addressable arrays to develop a device that performed automated electric-fielddriven immunocapture and DNA hybridization [111].
344 Jovanovich and Horn 10.6.2.3 Electrochemiluminescence assays
Immunomagnetic separations have been combined with electrochemiluminescence (ECL) detection [85]. ECL-based assays contain ruthenium labels, which emit light when electrochemically reduced. In ECL detection, tripropylamine can be oxided at the surface of an array of electrodes, and in turn it reduces the ruthenium, which then emits light. Background noise is reduced since the reaction activation is localized and controlled by the electrode. Several commercial systems are available, and over 50 immunoassays are available on one clinically aimed system. ECL has been further developed with microelectrodes manufactured using screen-printing of carbon inks onto microtiter plates. For biodefense, work in 1995 showed that the ECL detection of biotoxoids and 100 Bacillus anthracis spores in less than one hour [112]. An ECL assay was compared with fluorogenic chemiluminescence (FCL) for the detection of biological threat agents [113]. SEB at a concentration of 1 pg/mL has been detected in a range of matrices using ECL in a 30 min immunoassay and found to be significantly better than ELISA reactions [114]. Clostridium botulinum toxins A, B, E, and F were detected at about 100 pg/mL in clinical samples and food using IMS ECL detection, about the same sensitivity as ELISA, but with much more rapid time to results [115]. ECL immunoassays have also been used to detect ricin at 0.1 ng/mL [116]. In the future, ECL detection may well be integrated with microfluidics to produce fully integrated and sensitive laboratory and portable detection systems for biothreat agents. 10.6.2.4 Time-resolved fluorescence
Time resolved fluorescence uses lanthanide chelate labels that have very long fluorescent decay times and large Stokes shifts [85]. The secondary antibody can use a lanthanide such as Europium that fluoresces in an enhancer solution. The long decay time can lead to a very low background and sensitivity that are an order of magnitude better than traditional ELISAs, but with greater variability [117]. Commercial full scale systems are in use to detect biothreat agents including Franciscella tularensis, Clostridium botulinum toxin, and SEB with detection at a range of low pg/ml [118]. 10.6.2.5 Magnetic force assays
Immunoassays have been integrated with magnetic detectors to produce microfluidic systems that are being applied to biodefense and other fields.
Microfluidic Applications in Biodefense 345
Using this approach, the strength of intermolecular interactions can be measured by the force required to disrupt a bond when the target is attached to a magnetic bead [119-120]. The magnetic bead serves as the label which can be detected by microfabricated magnetoresistive transducers on microchips. Multiple analytes can be measured in less than 15 min in an array format with sensitivity close to ELISA [121]. Model spores and viruses can be detected at about 105 cfu/mL and 107 pfu/mL, respectively, while SEB was detected at 10 ng/mL. Multiple samples can be measured simultaneously and magnetic force is tolerant of different types of analytes. Multiplexed femtomolar detection of proteins from complex mixtures has been shown in a format that may be adapted for a handheld platform for both nucleic acid hybridization assays and immunoassays [122].
10.7 Proteomic Approaches Proteomic approaches for biodefense rely on identification of proteins and peptides to evaluate and characterize potential biothreat agents. Primary proteomic approaches for biodefense include separation of proteins and peptides by mass spectrometry platforms [123], two-dimensional gel analysis [124], and protein arrays. Proteomic approaches are being used to build a cyberinfrastructure of NIAID-funded centers that are applying these tools for biodefense to develop vaccines and proteomic targets [125]. Data from mass spectrometry, yeast two-hybrid (Y2H), gene expression profiles, X-ray and NMR for Bacillus, Brucella, Cryptosporidium, Salmonella, SARS, Toxoplasma, Vibrio and Yersinia, human tissue libraries, and mouse macrophages have been developed [126]. Microfluidics is being widely adapted to proteomic systems as upstream devices and nozzle systems for mass spectroscopy, as reviewed in [127128]. In essence, microfluidics is compatible with the low flow rates, small sample volumes, and microseparations that are required for mass spectrometry. Both microfabricated nozzles and HPLC devices are presently commercially available. The various developments are beyond the scope of this review, but are described in Chapter 3 of this book. For biodefense applications, protein profiling has been developed at Sandia National Laboratory into an autonomous microfluidics system combining microfluidics sample preparation modules with microchip gel electrophoresis [129]. This system is fully automated, with a total 10 min sample
346 Jovanovich and Horn
preparation and detection time. Sensitivity for B. subtilis spores was 16 agent-containing particles per liter of air. Protein arrays [130] containing either antibodies to different epitopes or to different proteins arrayed on a solid surface, have been applied to characterize and type biothreats. A protein chip for the ArrayTube platform was developed that uses a microtube-integrated protein chip that accomplishes detection using the classical sandwich assay and horseradish peroxidase colorimetric substrate. Immunoassays were developed for SEB, ricin, Venezuelan equine encephalitis virus, St. Louis encephalitis virus, West Nile virus, Yellow fever virus, Orthopox virus species, Franciscella tularensis, Yersinia pestis, Brucella melitensis, Burkholderia mallei and Escherichia coli O157:H7 [131]. Invitrogen has been developing its highdensity protein microarrays (ProtoArrays™) for detection of plague, smallpox, anthrax, and a number of hemorrhagic diseases, such as ebola and dengue fever. Phage display, where the probes on the array are selected from billions of clones, is a potentially powerful application of protein arrays and may be adopted for biodefense [132], as are peptide arrays.
10.8 Nucleic Acid Amplification and Detection Methods For nucleic acids following sample preparation, amplification technologies can be applied that greatly increase the signal. The best known and most used DNA amplification method is PCR [133]. PCR uses thermal cycling to exponentially amplify DNA using a thermally stable DNA polymerase. Each cycle of amplification doubles the amount of template, thereby exponentially increasing the amount of target, and can amplify from as little as a single copy of DNA. The specificity of the amplification is determined by the pair of primers that initiate the amplification. PCR has become a standard clinical and research technique for nucleic acid testing (reviewed in [134-136]) and for biodefense [137]. Many different variations of the basic PCR reaction have been developed, including qPCR, nested PCR, multiplexed PCR, and single nucleotide polymorphism PCR. qPCR has revolutionized the detection of specific DNA sequences in the laboratory, clinic, and field (reviewed in [138-139]). qPCR quantifies the amount of original target sequence using a fluorescently-labeled probe that acts as a reporter and detection system. In one version, qPCR employs a fluorescently–labeled probe that contains a quencher that suppresses the fluorescent signal. During the replication process, the signal becomes un-
Microfluidic Applications in Biodefense 347
quenched, emitting light. As the amplification progresses, the fluorescent signal increases. The presence and amount of target DNA in the sample can be determined from the cycle number when the signal increases over a threshold value, commonly called the cycle threshold (CT). Multiplexed reactions make it possible to detect multiple agents in each reaction vessel. The amount of DNA produced can also be measured non-specifically using DNA intercalating dyes and other (specific) probes such as molecular beacons. qPCR has specificity and sensitivity equivalent to PCR, while simultaneously amplifying, detecting, and quantifying the original DNA target in a single, contained reaction. 10.8.1 PCR and qPCR detection of pathogens for biodefense PCR and qPCR are the most widely used methods to detect and identify biowarfare agents by identifying target DNA sequences in the laboratory and field [140]. Multiplexed PCR assays for virulence factors on two plasmids, pXO1 and pXO2, in B. anthracis were able to distinguish it from closely related strains [141]. Highly specific assays can identify B. anthracis using pXO1, pXO2, protective antigen (pagA), and capsular protein B (capB) [142] are commercially available for clinical samples [143144]. B. anthracis has been detected from soil samples at 10 spores per mL [145], in aerosols [146], and in food [147]. qPCR detection assays have been multiplexed to detect four biothreat agents, Y. pestis, F. tularensis, B. anthracis and B. mallei, simultaneously using molecular beacons complementary to conserved 16S rRNA targets [148], and with minor groove binding probes with a sensitivity of 1 fg of target with no cross reactivity [149]. Melting point curves combined with the multiplex amplification helped distinguish B. anthracis from Y. pestis and Leishmania [150], and between members of the B. cereus group [151]. The performance of three commonly used qPCR instruments was compared and generally comparable limits of detection, sensitivity, and specificity were found [152]. The GeneXpert system (Cepheid) is a standard instrument for semi- automated sample preparation and qPCR (using the SmartCycler as the qPCR platform). Its performance has been evaluated and the incorporation of the nucleic acid purification component of the sample preparation has been credited with its 1,000-fold improved detection over the SmartCycler alone for B. anthracis Sterne spores due to removal of inhibitors and concentration of the sample [153].
348 Jovanovich and Horn
10.8.2 Miniaturized and Microfluidic PCR Miniaturized PCR has been an area of intense development with devices developed that use microchips and capillaries to perform thermal cycling in stationary formats and continuous flow regimes. The advantages of miniaturized PCR devices are more rapid assay times, low reagent consumption, and potential integration with upstream sample preparation modules. Miniaturization of PCR, including discussion on the types of common designs, issues with surface chemistries that can inhibit PCR, coating procedures, and heating strategies has been reviewed [154-155]. Miniaturized PCR was initially performed in glass capillary tubing. Capillaries are off-the-shelf items that are suitable for sub-microliter reactions. Wittver first demonstrated (in 1989) PCR amplification of DNA in sealed capillary tubes using a hot air cycler [156]. This basic design became the Rapid Cycler (Idaho Technologies), the forerunner to RAZOR, in use today for biodefense applications. Sample volumes were reduced to 1-10 uL with cycling times of less than 15 min [157]. A medium-throughput automated capillary sample preparation system that processed 1,000 one uL samples per day was developed [158]; the Acapella-1K (U. Washington, Seattle) utilized a mechanical capillary handling system, air cycling, and a piezoelectric reagent dispenser [159]. Jovanovich and co-workers developed a 500 nL sample preparation system also using an air cycler with a 384-channel capillary cassette (Fig. 10.2). This system used 150 um ID capillaries without a coating and standard PCR conditions (except for elevated Mg+2 concentrations). DNA samples were forced onto the surface of the glass capillary using chaotrophic agents, the DNA-coated capillary was then evacuated, and 500 nL of PCR reagents re-filled the tube by capillary action. PCR reaction efficiencies were equivalent to full volume reactions. Capillaries have many applications in microfluidics but are currently more difficult to integrate with other functions than microchips. The integration requires robust microfluidic connectors.
Fig. 10.2 Capillary cassette with 384 capillaries, each with a volume of 500 nL.
Microfluidic Applications in Biodefense 349
The first miniaturized PCR reactions in a microchip were performed in an etched silicon wafer by Northrup and coworkers, reported in 1993 [160]. Wilding et al. reported successful cycling in 5 and 10 uL silicon reactors cycled by an external Peltier device in 1994 [161]. PCR reactions have been performed in 87 nL in silicon devices [162], which was expanded on by Belgrader et al. by addition of capillary electrophoresis on microchip [163]. Belgrader and coworkers further reported PCR amplification in seven minutes [164]. The first miniaturized thermal cycler with real time detection was from Northrup and colleagues at LLNL and used silicon based reaction chambers with diode detectors and integrated heaters [165], this same group devised a portable battery-powered unit [166]. Many other reports of qPCR devices have followed [167-169]. Some of the challenges of microfluidic qPCR are control of liquid positioning, bubble formation, sealing chambers using microfluidic valves, and surface interactions. Liquids and targets can be manipulated by precise positioning by micropumps, by capture onto a surface, or by exploiting beads and solid phase chemistries. Microvalves can restrict the liquid solutions to proper chambers. One group has simultaneously sealed the reaction chamber with valves that serve as the macro-to-micro interface [170]. For poly(cyclic olefin) plastic fabrication, gel valves were used to confine amplicons before integrated on-chip gel electrophoresis [171]. Bubble formation has been found to be primarily due to surface wetting properties of the chamber [172]. Surface interactions can be minimized by improving the surface chemistry or by increasing channel dimensions to over 125 microns. 10.8.3 Heating and cooling approaches A challenge for microfluidics is improving heating and cooling rates while maintaining good uniformity. One approach has been to microfabricate resistive heaters and sensors directly onto microchips. This has been used to achieve 20o C/s heating and 10o C/s cooling rates to detect upper respiratory tract infection microorganisms in 15 min [173]. A second approach, pioneered by the Landers group, used infrared heating to rapidly heat water in fluidic channels [174-175]; cycle times as low as 17s were achieved with successful PCR amplification and cycle sequencing. IR driven PCR was integrated on a microchip with upstream solid-phase extraction of DNA using silica beads in sol gel to isolate DNA from an anthrax spore-
350 Jovanovich and Horn
spiked nasal swab; amplified target DNA sequences were detected on-chip with a total processing time of 25 min [176]. A third approach is flowthrough devices with fixed-temperature zones. Control of temperature and uniformity is simplified, and power consumption is low since the zones stay at a constant temperature. An added advantage is that cycling times can be reduced due to the elimination of temperature ramping. The challenge to this approach is to reduce surface interactions which can degrade PCR reaction performance. This can be accomplished using surface coatings, by controlling surface to volume ratios, or by using emulsions. Soper’s group has built flowthrough devices combining flowthrough PCR in a polycarbonate chip and a PMMA chip using detection with a ligase assay [177]. Reverse transcription and PCR have been integrated in a flow through glass microchip with 55 um channels integrating different temperature zones; 30 cycles of PCR amplification were achieved in 6 min [178]. A ferrofluidic actuator has also been used to move a PCR sample plug rapidly between temperature zones [179]. A hybrid chip with silicon and PMMA has performed high speed PCR with three heated zones [180]. PCR amplification in small volumes has also been accomplished in microarray formats. The target DNA is spotted onto a glass slide that has covalently attached primers, enabling amplification of bacterial target DNA [181]. This approach has been applied to identify bacteria using amplification of rDNA sequences [182]. The Solexa DNA sequencer uses a similar process to amplify clusters of PCR products from a single template molecule on a surface. 10.8.4 Miniaturized PCR and qPCR for biodefense One of the first miniaturized real-time PCR instruments was developed for biodefense applications. The Advanced Nucleic Acid Analyzer (ANAA), devised at LLNL, used an array of 10 silicon reaction chambers with thinfilm resistive heaters and solid-state optics to rapidly test samples for simulants of biothreat agents with detection limits of 100 to 1,000/mL [183-184]. A compact version of a qPCR instrument with a notebook computer, two reaction modules with integrated four-color fluorescence detection was developed and shown to detect Bacillus spores [166]. A handheld device, the Handheld Advanced Nucleic Acid Analyzer (HANAA), also developed at LLNL, used plastic reaction tubes with silicon and platinum-based thermal cycler units and two light emitting diodes
Microfluidic Applications in Biodefense 351
to detect bacteria including B. anthracis Ames [185]. The LLNL group, working with Sandia, went on to develop the APDS, which performs continuous monitoring with a multiplexed immunoassay trigger and confirmatory qPCR assays. This has been demonstrated for aerosolized B. anthracis, Yersinia pestis, Bacillus globigii, and botulinum toxin [186]. LLNL and Sandia also developed a “Biobriefcase” device with a smaller footprint that has multiplexed, autonomous detection with immunoassays, toxin assays, and qPCR. The Biobriefcase incorporated inhibitor removal and concentrated the sample by mixing aerosol-collected liquid with a chaotrophic agent prior to silica bead bed purification [187]. The LLNL group has also developed a 10-plex PCR amplification with hybridization to beads upstream of a flow cytometry readout; 1000 samples were processed in 8 hours [188]. The LightCycler (Idaho Technology) has been adapted into a “ruggedized” Advanced Pathogen Identification Device, RAPID, which has been used for field analysis of bioagents and pathogens [189]. Y. pestis was detected at about 20 genome equivalents in 75 min using this system. The latest innovation from LLNL is amplification in emulsion of single copy DNA in a lab-on-a-chip format; the system produces 10 pL droplets that can perform qPCR with thresholds exceeded significantly earlier than conventional instruments [190]. In addition to detecting DNA targets, biodefense needs require that RNA viruses be detected. Detection of two RNA-based viruses—Dengue virus and enterovirus 71 has been demonstrated with a PDMS microchip that integrates reverse transcriptase and PCR amplification [191]. Parallel reverse transcriptase-PCR assays have been performed in 450 pL and shown to detect as little as 34 copies [192]. A fully integrated handheld device using isothermal nucleic acid sequence-based amplification (NASBA) [193], which specifically amplifies RNA from primers, was built and shown to have comparable results to laboratory instruments. 10.8.5 Other Nucleic acid amplification methods In addition to PCR there are a number of other nucleic acid amplification technologies that are being adapted for biodefense: strand displacement amplification (SDA), loop-mediated isothermal amplification (LAMP), exponential amplification reaction (EXPAR), and rolling circle amplification (RCA). SDA uses the primer-directed nicking activity of a restriction enzyme and the strand displacement activity of an exonuclease-deficient polymerase to amplify DNA [194]. SDA can achieve 108 to 1010 amplification in about 15 min [195]. Fluorescence resonance energy transfer
352 Jovanovich and Horn
(FRET) probes can be incorporated and real time instruments for the clinical laboratory are in commercial use [196-197]. EXPAR is a DNA amplification reaction that produces short (8-22 nucleotides) in a linear or exponential amplification in isothermal homogeneous assays [198]. EXPAR can be extremely rapid with amplification of greater than 106. Both real time [199] and end point formats have been developed. EXPAR has been applied to biodefense applications as part of the BAND project and in other projects as a quick, specific amplification technique. LAMP is an isothermal DNA amplification method that relies on autocycling strand-displacement DNA synthesis. Four sets of primers are used with the thermostable Bacillus stearothermophilus (BST) DNA polymerase that has high strand displacement activity [200]. The reaction produces a white precipitate, magnesium pyrophosphate, which can be easily detected as an indicator of a positive amplification reaction [201]. The specificity and sensitivity are competitive with PCR, and in some cases more sensitive than nested PCR [202]. RCA is an isothermal method that uses Phi29 DNA polymerase to amplify circular DNA [203]. Phi29 DNA polymerase is a single subunit with excellent processivity and can amplify >107 fold [204]. RCA is widely used for whole genome amplification, for scarce material, and can nonspecifically amplify trace amounts of DNA. RCA is routinely used at genome centers preparing template for sample preparation for DNA sequencing [205]. RCA has also been used for cell free cloning of genomic DNA that might be lethal to cells [206]. RCA is a powerful tool for forensic and biodefense applications.
10.9 Microarrays DNA microarrays have many potential applications in biodefense. DNA microarray technology is a widely used powerful technique that uses large arrays of microspots of DNA on a solid support or beads to detect complementary DNA or RNA products from a sample. (Protein arrays were briefly discussed in the Proteomics section). For RNA samples, amplification of RNA is commonly done by first performing reverse transcriptase to create DNA from RNA, and then the resulting DNA can be amplified using standard DNA amplification methods such as PCR, or whole genome amplification, or by in vitro transcription followed by another cycle of re-
Microfluidic Applications in Biodefense 353
verse transcriptase. Fluorescent labeling can be introduced into the sample at various steps in the process. Typically, tens of thousands of spots, each containing a unique sequence, are interrogated in a single experiment with fluorescent detection. The data can represent a fingerprint of the transcriptional state of an organism (e.g. biothreat agent, or of the response of a human potentially infected with the agent), identify DNA sequences present in an organism, or resequence organisms [207-208]. The strength of the microarray platform is the depth of characterization. The 10,000s of analytes measured on a single microarray slide can generate massive amounts of data. In the future, DNA microarrays may be displaced or challenged by digital gene expression methods using next-generation DNA sequencing to produce 100,000’s or more sequences from a sample or single mRNA detection and enumeration strategies (e.g., Nanostring). Microarrays have been combined with PCR amplification to identify and genetically discriminate B. anthracis from closely related bacterial species from the B. cereus group and determine if the strains harbor plasmids [209]. DNA microarrays can potentially detect multiple pathogens in a single sample. The FDA has developed a microarray, FDA-1, to screen for several food pathogens and virulence factors, including SEB [210]. SAIC developed its Phase I BAND project based upon a fully automated system that would collect air samples and then analyze them for pathogens using microarrays. After collection and filtration, reverse transcriptase and PCR amplification were used before hybridization to a DNA microarray in a cartridge for ten min. The data was then analyzed for fingerprints that indicated the presence of threat agents. 10.9.1 Microarrays and microfluidics Microarray sample preparation is a complicated, multistep process that is dependent on the variability of individual operators. Microfluidics, with its potential to automate and integrate processes, has been applied to simplify and standardize sample preparation (reviewed in [211]). This seminal work was performed by Anderson and colleagues at Affymetrix, where a miniaturized integrated microdevice was developed to prepare samples by accessing 10 reagents and performing 60 automated operations before performing hybridizations to a microarray. Anderson and coworkers elegantly demonstrated a plastic miniaturized sample preparation system for microarray sample preparation and analysis by hybridization, and showed the detection of mutations in the HIV genome from serum samples [212]. The advantages of microfluidics for microarray biodefense applications are in-
354 Jovanovich and Horn
tegration of sample preparation steps, potential reductions in reagent cost, ease of use, and decreased hybridization times. The Soper group has integrated microarrays with microfluidic technology in plastics. A microarray was fabricated in PMMA using UV exposure of the polymer surface, coupling of amine-terminated oligonucleotide probes to the surface, and washing the surface [213]. The hybridization and allelespecific ligase detection reactions were performed in a polycarbonate flowthrough biochip. The system could screen for mutations in 20 min. Work by Liu and coworkers at Motorola and then Combimatrix has developed biochips that integrate sample preparation, PCR, and microarray detection [214]. Electrochemical pumps were used with paraffin-based single-use microvalves to regulate flow. Detection of pathogens from whole blood [214], identification of influenza virus [215], and gene expression from a cell line [216] were shown. A microelectronic array system was developed by Nanogen for microarray testing [217]. This system employs dielectrophoresis as a sample preparation method and the hybridization of nucleic acid to probes attached to electrodes is accelerated by application of an electrical potential. The electrodes are covered with a hydrogel. Detection is performed with fluorescence probes [218].
10.10 Microelectrophoresis and Biodefense Electrophoresis is a powerful separation technology with many biodefense applications. The advent of capillary electrophoresis (CE) and the multichannel capillary array electrophoresis (CAE) propelled the applications of electrophoresis by increasing separation speeds, automating the analysis, and improving data reliability. Most notably, CAE technology was partially responsible for the early completion of the Human Genome Project and has become the separation method of choice for most nucleic acid applications. Several PCR-based methods for detailed laboratory characterization of microorganisms have been developed that rely on electrophoretic separation. In this section, we first review several of the separation based typing methods for bacteria identification and discrimination, describe the application of microfluidics on microchips to microelectrophoresis, and discuss
Microfluidic Applications in Biodefense 355
applications of DNA sequencing in a microfluidic platform to biodefense [219]. VNTR loci analysis [220] is a powerful laboratory tool for identifying bacteria, humans, and other organisms. Analogous to microsatellite genotyping, VNTR PCR amplification detects regions where genetic drift has created variable numbers of tandem repeats inserted into the genome, plasmids, or other extra-chromosomal elements. VNTR has been used to identify Y. pestis [221], Mycobacterium tuberculosis [222], and numerous other organisms. MLVA [223] extend VNTR to assay multiple alleles and provide a fingerprint, analogous to forensics identification by microsatellite DNA analysis. Keim and co-workers have identified a set of eight VNTR regions that are diagnostic for B. anthracis and characterized 426 B. anthracis isolates into 89 distinct genotypes [223]. MLVA was used to subtype the anthrax strains from the bioterrorist attack in 2001 within eight hours of receiving isolates [224]. MLVA has been applied to type closely related Bacillus strains but at times required additional information for confirmatory determination of B. anthracis [225]. MLVA types of Bacillus anthracis could be further differentiated by single-nucleotide repeats [226]. AFLP is a PCR-based method that can fingerprint microbes [227], type organisms, and identify phyllogenetic relationships. AFLP is rapid—it employs a relatively simple multi-step workflow with standard reagents regardless of the organism typed. AFLP has been applied to differentiate a wide variety of microorganisms at the subspecies level [228], including B. anthracis [229] and is widely used to classify plants, yeasts, and other organisms. The AFLP process begins with a two-enzyme restriction digest followed by ligation of fluorescently labeled restriction half-site adapters to reconstruct the restriction site at the end of the fragments and serve as PCR primers. Two or three degenerate nucleotides on the end of the adapter reduce the complexity of the amplified products during the high stringency PCR amplification. By selection of proper restriction enzyme pairs, fluorescently labeled fragments in the range of 100 to 1,000 bases can be produced, and then are separated and detected by capillary electrophoresis. The resultant pattern is a fingerprint of the test organisms’ genomic and extra-chromosomal restriction patterns. A post-labeling fluorescent method using dye-terminator chemistry can visualize both RFLP and AFLP products [230]. cDNA-AFLP can also be performed to visualize the gene expression pattern of an organism [231].
356 Jovanovich and Horn
Subtle strain-to-strain variations can be characterized by single nucleotide polymorphisms (SNPs). SNP typing of intragenic spacers in the 16S-23S region has been shown to differentiate closely related Bacillus strains including B. cereus from B. anthracis [232]. The identification of the specific strain of a biothreat organism can help microbial forensics trace the origin and determine whether multiple incidents were caused by release of identical organisms and therefore share a common origin. To determine strain variations, the genomic sequence can be determined by DNA sequencing. Following the September 11th attacks, the CDC sequenced the 16S rRNA gene to definitely identify B. anthracis from culture-negative clinical specimens of patients with confirmed anthrax [233]. In other studies, 183 16S rRNA and 74 23S rRNA sequences for all species in the B. cereus group showed disagreement with phenotyping clustering, but by utilizing rRNA together with gyrB sequences these workers could discriminate between groups [234]. Ruppitsch and coworkers showed for a wider variety of strains that sequencing of the 16S rRNA gene is not always sufficient, but that additional sequencing of intragenic sequences can increase resolution and thus differentiate bioagents [235]. 10.10.1 Microelectrophoresis technologies CAE is based upon separation in a capillary, itself a type of microdevice. In this review, we use microelectrophoresis to connote separations on microchips. The interested reader is referred to reviews [236-238] for details of chip construction, coatings, operation, and equipment. Most commonly, glass microchips are employed but plastic devices have also been developed [239]. Microchips have the potential to simultaneously separate hundreds of samples in minutes. Microchips typically consume only picoliters of samples and thus can be well matched to microscale sample preparation volumes. Fluorescently labeled amino acids [240], DNA restriction fragments [241243], PCR products, short oligonucleotides [244], and sequencing ladders [245] have been separated by microchip capillary electrophoresis [246]. The analyses are extremely rapid, from less than a minute for oligonucleotides to less than 20 minutes for DNA sequencing [247]. We note that next generation microfluidic sequencing-by-synthesis [248], microfluidic pyrosequencing [249], sequencing by ligation, and nanopores may, in the future, have biodefense applications for detecting genetically modified organisms and digital gene expression.
Microfluidic Applications in Biodefense 357
Interfacing upstream microfluidic sample preparation with microscale separations is challenging. The classic Twin-T injector defines roughly a 100 to 500 picoliter volume in a short sample plug (100 to 500 um) in the separation channel with electrokinetic loading [250]. This is only a small fraction of the sample, even if the prepared sample is only 100 nL: for biodefense applications this means losing orders of magnitude of potential sensitivity during the analysis step. This can be ameliorated with isotachaphoresis, field amplified stacking, DNA binding, and other methods that locally increase the sample concentration. Isotachaphoresis concentrates samples between leading and trailing electrolytes and can increase detection limits by 50-fold or more for DNA [251-252], and stack up to a million-fold [253]. Field amplified sample stacking, routinely used in capillary electrophoresis, has been adapted to microchip electrophoresis using pressure driven flows to move low osmotic strength samples into position for stacking on the column. Concentration effects range from 65-fold increase in signals [254] to 180-fold for model compounds [255]. Separations on microchips are most developed for DNA separations. In 1998, 96 hemochromatosis samples were genotyped in less than 8 min by microelectrophoresis on a microchip [256]. Locus-specific, multiplex PCR products specific for deletions causing Duchene/Becker muscular dystrophy have been separated on a silicon-glass microchip, and T-cell receptorgamma genes and immunoglobulin heavy chain gene on glass microchips [257]. Fluorescently-labeled CTTv PCR samples and short tandem repeats have been analyzed on single-channel CE microchips [258]. The Mathies laboratory genotyped 384 hemochromatosis samples in less than 6 min, with detection by a four-color rotary confocal fluorescence scanner [259]. DNA sequencing on microchips was first performed in 1995 in the Mathies’ lab [260]; fluorescently labeled DNA sequencing fragment ladders were separated on glass microchips with a denaturing polyacrylamide sieving matrix and readlengths of about 200 bases obtained in 15 min. Mathies’ lab later reported readlengths of about 500 bases in about 30 min in single channel [261]. Jovanovich’s group reported the first multichannel DNA sequencing in an array of 16 microchannels with readlengths of 450 bases in 15 min in an automated instrument [247]. High throughput DNA sequencing in 96 channels obtained readlengths of 430 high quality bases [262]. To achieve longer readlengths, microchips with 40 to 50 cm separation channels were constructed and readlengths of 600 to 800 bases reported with plates containing up to 384 channels [263].
358 Jovanovich and Horn
Separations on plastic microchips have also been achieved. One of the first microelectrophoresis devices, in 1993, used a twin T injector on a PDMS substrate to separate proteins and DNA [289]. Plastic substrates offer the advantage of potential low cost fabrication once high production volumes are achieved, but have been plagued by higher background fluorescence signals than glass or quartz devices and require specialized surface coatings [264]. While a detailed review is outside the scope of this article, injection molded microelectrophoresis devices have achieved separation of DNA fragments [265-266] and DNA sequencing [267]. Hot embossed PMMA with IR detection was shown to sequence DNA out to 450 bases in linear polyacrylamide [268]. For biodefense, the breakdown products of G-type and V-type nerve agents can be assayed by CE and microelectrophoresis [269]. Wang et al. [270] combined microchip electrophoresis with derivatization and electrochemical detection of thiol-containing degradation of V-type nerve agents at micromolar concentrations in less than four min. A microChemLab portable device, developed at Sandia National Laboratory, with a fused silica microfluidic separation chip, a miniature LIF detector, and high voltage power supplies [271] can assay proteins and toxins at nanomolar concentrations [272].
10.11 Integrated lab-on-a-chip systems and biodefense The achievement of working devices for both microfluidic based sample preparation and analysis has created the possibility for complete system integration to meet biodefense needs. For robust biodefense applications, sample preparation and analysis must be integrated either directly in a monolithic format or by microfluidic connections. As described above, for DNA microarrays sample preparation has been integrated with analysis, as first shown by Anderson, and brought to commercial product by Combimatrix. Today, one of the most advanced areas for complete system integration has been the integration of upstream sample preparation reaction steps with microelectrophoresis on microchips. Microchip and capillary-based analysis systems have the advantages of high-resolution separations with extremely fast separation times, automation, and nanoliter-scale consumption of reagents. Integration of PCR reactions in microfabricated devices with microelectrophoresis was first demonstrated in 1996 in collaboration be-
Microfluidic Applications in Biodefense 359
tween Northrup’s LLNL group and the Mathies laboratory; PCR amplification was performed in a silicon device mounted on a glass separation device and complete analysis was achieved in 20 min for a cloned human gene and 45 min for a bacterial genomic target [273]. Early work exploited the small volumes in capillaries to prepare samples and analyze them by capillary electrophoresis. Swerdlow and colleagues developed an automated prototype with a sample loop in an air cycler coupled to a separation capillary [274]. A fully integrated miniaturized integrated microsystem using capillaries was shown to perform all steps including PCR amplification in 125 nL and separations starting from buccal cells [275]. Integration of PCR reactions in capillaries with separations in capillaries has been shown to be effective for DNA typing from blood [276] and other materials for human and viral targets [277]. PCR reactions and DNA cycle sequencing in capillaries and capillary separations have been fully integrated to create a microvolume system with readlengths of 257 bases in 4 hrs from human DNA [278].
360 Jovanovich and Horn
Fig. 10.3 Integrated sample preparation, clean-up, and analysis for DNA sequencing on a microchip, from [283]. Bioprocessor components. (A) Photograph of the microdevice, showing one of two complete nucleic acid processing systems. (Scale bar, 5 mm.) B–F correspond to the following component microphotographs. (B) A 250-nl thermal cycling reactor with RTDs. (Scale bar, 1 mm.) (C) A 5-nl displacement volume microvalve. (D) A 500-um-diameter via hole. (E) Capture chamber and cross injector. (F) A 65-um-wide tapered turn. (Scale bars, 300 um.) All features are etched to a depth of 30 um. Copyright (2006) National Academy of Sciences, U.S.A.
Microfluidic Applications in Biodefense 361
PCR reaction chambers have now been monolithically integrated with microelectrophoresis to create fully integrated microsystems that perform PCR on the same device as the separation. Burns et al. integrated sample preparation, gel electrophoresis, and on-chip detection for low resolution separation of DNA fragments [279]. The Mathies lab showed PCR amplification and microelectrophoresis separation to determine sex of humans from DNA in 15 min [280]. A portable system with 200 nL PCR reactors, solid state lasers, pneumatically actuated on-chip micropumps, and microelectrophoresis was able to detect 2-3 E. coli or S. aureus cells in less than 10 min including determining drug resistance [281]. The Mathies lab has extended their DNA sequencing on microchips upstream to include integrated nanovolume sample preparation and purification. First, nanoscale PCR reactions were combined with capillary electrophoresis analysis [282]. Then, sample preparation, cleanup, and DNA separations were combined to integrate DNA sequencing, as shown in Fig. 10.3, [283]. A 250 nL cycle sequencing reaction was performed on microchip, and then moved by micropumps onto an acrylamide gel capture matrix with an oligonucleotide hybridization probe to capture the target DNA in a 60 nL capture chamber. The affinity capture removes template DNA, desalts, and pre-concentrates the sample for microchip electrophoresis for higher injection efficiencies. The sample is then electrophoresed into an injector and separated on microchip. Readlengths up to 556 bases were obtained from 1 fmol of template. The key to the integration was pneumatically actuated microvalves and micropumps [288]. The Landers laboratory has completely integrated PCR, sample cleanup, and capillary electrophoresis on microchips, as shown in Fig. 10.4. A device with PCR reactions, IR mediated heating, pneumatically actuated onchip micropumps, and microelectrophoresis has achieved amplification and good separations in less than 12 min [284]. They have detected of B. anthracis from whole blood of asymptomatic mice and Bordetella pertussis from nasal aspirate of a patient using on-chip nucleic acid purification with a 550 nL PCR reactor coupled to microelectrophoresis analysis with control by pneumatic microvalves [285].
362 Jovanovich and Horn
Fig. 10.4 Fully integrated device for biodefense detection, from [285]. Images of the device. (a) (Scale bar, 10 mm.). Domains for DNA extraction, PCR amplification, injection, and separation are connected by channels and vias. SPE reservoirs are labeled for sample inlet (SI), sidearm (SA), and extraction waste (EW). Injection reservoirs are labeled for PCR reservoir (PR), marker reservoir (MR), and sample waste (SW). Electrophoresis reservoirs are labeled for buffer reservoir (BR) and buffer waste (BW). The flow control region is outlined by a dashed box. (b) Schematic of flow control region. Valves are shown as open rectangles. (c) Device loaded into the manifold. (d) Intersection between SI and SA inlet channels, with the EW channel tapering to increase flow resistance. (Scale bar, 1 mm.) (e) Image of PCR chamber with exit channel tapering before intersecting with the MR inlet channel. (Scale bar, 1 mm.) (f) Image of cross-tee intersection. (Scale bar,1 mm.). Copyright (2006) National Academy of Sciences, U.S.A.
Microfluidic Applications in Biodefense 363
10.11.1 Full microfluidic integration for biodefense Fully integrated ‘industrial strength’ microfluidics is being applied to Homeland Security and biodefense in the BAND program by several groups including Microfluidic Systems Inc, IQuum, and US Genomics. IQuum (www.iquum.com) is developing a Liat™ “detect-to-treat” system. The system has a disposable cassette containing reagents and equipment that contains air sampling, sample preparation, and real time detection using PCR in a “lab-in-a-tube” [286]. The sample moves through segmented tube sections containing the reagents using peristaltic pumps and moves back and forth between temperature zones to amplify the DNA. US Genomics (www.usgenomics.com) has taken a very different approach. Single stranded DNA is prepared from aerosol samples and a set of fluorescently labeled probes added. The probes bind to their homologous targets, if present. The labeled DNA is then moved through a narrow channel one molecule at a time where it is interrogated by laser-induced fluorescence. The resulting pattern of binding sites is a fingerprint that can identify known and unknown agents from a database [287]. One advantage of the approach is the analysis of individual molecules is done without apriori knowledge of the sequence. This potentially enables detection of genetically engineered or previously unknown pathogens. In the future, next generation DNA sequencers, mass spectroscopy, and DNA microarrays will provide additional solutions to begin to address genetically engineered organisms.
10.12 Summary and Perspectives Microfluidics will become an increasing important element of the future biodefense portfolio. As shown in this and other chapters in this volume, microfluidic components are available in many different shapes and formats: proof-of-concept of just about any imaginable type of sample extraction, lysis, pre-separations, sample preparation, assay, separation, and detection component has been demonstrated in research settings, at universities, research facilities, national laboratories, and in industry. The missing component is the integration of the many parts into complete working systems that are robust, manufacturable, and maintainable. Complete integration of complex workflows has proved elusive in full volume devices and in microfluidics. To fully integrate a microfluidic system for biodefense requires taking a sample of several hundred microliters or
364 Jovanovich and Horn
larger and processing it completely to yield an answer. The chemical and biochemical workflow usually has many steps that all have to be developed, tested, and integrated. In addition, of course, surface chemistries, temperature, and other variables must be controlled. The work described in this review has shown the promise for microfluidics for many types of genomic and proteomic devices. Fully microfluidic integrated devices that can input samples and perform sample preparation and analysis are just beginning to appear, as are field portable microfluidic genetic analyzers. While there are some examples of success at the research and prototype levels, there are only a few fully integrated solutions and fewer commercial successes. Two main microfluidic elements are needed for full integration. The first element is an appropriate level of control of the microfluidic volumes. This typically requires microfluidic valves, a pumping mechanism, and routers to move, mix, and split, and aliquot liquids. Single use valves enable many applications in a disposable cartridge, while programmable valves that enable multiple uses are required for environmental monitoring and reusable applictions. Pneumatically driven programmable microvalves are only now becoming more mature and are in use in academia and industry to control fluid flows and mixing. These microvalves [288] have proved invaluable in the integration by the Mathies and Landers groups, and in industry to integrate processes onto monolithic microdevices. The second enabling element will be to connect different microfluidic devices together using standard connections. In microelectronics, USB, Firewire, and other standard connectors allow devices to interconnect by defining the interface and the protocols. Microfluidics now needs to establish standards for connections that will enable ‘best in class’ microfluidic devices to work together in a ‘plug and play’ manner. A symposium of industry and academia with government representation, including DoD, DHS, and NIST, to discuss achievable initial standards would be invaluable and may serve to begin the path towards interconnectivity for microfluidics for biodefense and other applications. In this regard, the physical dimension of spacing for connectors needs to be standardized, and standard protocols for transfer of materials developed. The future for microfluidics will be bright as individual steps are optimized and integrated. Future microfluidic systems will connect directly to ‘real world’ samples and fully integrate upstream sample concentration and analysis in a single autonomous device. The full integration of micro-
Microfluidic Applications in Biodefense 365
fluidic processes will enable man-portable and then hand-held biodefense devices. Eventually, if biothreats become pervasive, microfluidic home and business security devices akin to smoke detectors may provide the massive sampling capability needed to detect to warn the public. The public and the biodefense community await the transformation that complete microfluidic integration of biodefense detection can bring to increase the biosecurity of the world.
References 1. Boscarino JA, Figley CR, Adams RE (2003) Fear of terrorism in New York after the September 11 terrorist attacks: implications for emergency mental health and preparedness. Int J Emerg Ment Health 5(4):199-209. 2. Venkatesh S, Memish ZA (2003) Bioterrorism—a new challenge for public health. Int J Antimicrob Agents 21(2):200-6. 3. Heller, MB, ML. Bunning, ME. France, DM Niemeyer, L Peruski, T Naimi, et al. (2002) Laboratory Response to Anthrax Bioterrorism, New York City, 2001. Emerg Inf Dis 8:1096-1102. 4. Bennett RL (2006) Chemical or biological terrorist attacks: an analysis of the preparedness of hospitals for managing victims affected by chemical or biological weapons of mass destruction. Int J Environ Res Public Health 3:67-75. 5. Rebmann T, Carrico R, English JF (2007) Hospital infectious disease emergency preparedness: a survey of infection control professionals. Am J Infect Control 35:25-32. 6. Alexander GC, Larkin GL, Wynia MK (2006) Physicians' preparedness for bioterrorism and other public health priorities. Acad Emerg Med 13(11):1238-41. 7. Williams, M (2006) The knowledge. Technology Review. 109:44-53. 8. United States General Accounting Office, (20000 Report to Congressional Requesters, Biological Weapons: Effort to Reduce Former Soviet Threat Offers Benefits, Poses New Risks. 9. Henderson DA (1999) The looming threat of bioterrorism. Science 283(5406):1279-82. 10. Petro JB, Carus WS (2005) Biological threat characterization research: a critical component of national biodefense. Biosecur Bioterror 3(4):295-308. 11. Cirino NM, Musser KA, Egan C (2004) Multiplex diagnostic platforms for detection of biothreat agents. Expert Rev Mol Diagn 4(6):841-57. 12. Klietmann WF, Ruoff KL (2001) Bioterrorism: implications for the clinical microbiologist. Clin Microbiol Rev. 14(2):364-81. 13. Hamilton MG, Lundy PM (2007) Medical countermeasures to WMDs: defense research for civilian and military use. Toxicology 233(1-3):8-12. 14. Russell PK (2007) Project BioShield: what it is, why it is needed, and its accomplishments so far. Clin Infect Dis 45 Suppl 1:S68-72.
366 Jovanovich and Horn 15. Shea DA, Lister SA (2003) The BioWatch Program: Detection of Bioterrorism. Congressional Research Service Report No. RL 32152. 16. www.milnet.com/wh/DoHS/BioWatchFactSheetFINAL.pdf 17. Comments made in "Meeting Minutes," CDC Information Council, February 27, 2003, found online at http://www.cdc.gov/cic/minutes/CIC%20minutes%202-27-03.pdf. 18. DHS’ Management of BioWatch Program, OIG-07-22, January 2007. 19. Hearing Before The Subcommittee On Technology And Innovation Committee On Science And Technology House Of Representatives One Hundred Tenth Congress First Session The “Department Of Homeland Security's R&D Budget Priorities For Fiscal Year 2008” March 8, 2007, Serial No. 110–8, 33– 611PS 2007. Available via the World Wide Web: http://www.house.gov/science. 20. Koplan J (2001) CDC's strategic plan for bioterrorism preparedness and response. Public Health Rep 116 Suppl 2:9-16. 21. Ashford DA, Kaiser RM, Bales ME, Shutt K, Patrawalla A, McShan A, Tappero JW, Perkins BA, Dannenberg AL (2003) Planning against biological terrorism: lessons from outbreak investigations. Emerg Infect Dis 9:515-9. 22. Crutchley TM, Rodgers JB, Whiteside HP Jr, Vanier M, Terndrup TE (2007) Agroterrorism: where are we in the ongoing war on terrorism? J Food Prot 70:791-804. 23. Cupp OS, Walker DE 2nd, Hillison J (2004) Agroterrorism in the U.S.: key security challenge for the 21st century. Biosecur Bioterror 2:97-105. 24. Niemeyer D. JBAIDS Team. (2004) Improving Laboratory Capabilities for Biological Agent Identification. J of Homeland Security March. 25. Lim DV, Simpson JM, Kearns EA, Kramer MF (2005) Current and developing technologies for monitoring agents of bioterrorism and biowarfare. Clin Microbiol Rev 18:583-607. 26. Jannes G, De Vos D. (2006) A review of current and future molecular diagnostic tests for use in the microbiology laboratory. Methods Mol Biol 345:121. 27. Peruski LF Jr, Peruski AH (2003 ) Rapid diagnostic assays in the genomic biology era: detection and identification of infectious disease and biological weapon agents. Biotechniques 35:840-6. 28. Ivnitski D, O'Neil DJ, Gattuso A, Schlicht R, Calidonna M, Fisher R (2003) Nucleic acid approaches for detection and identification of biological warfare and infectious disease agents. Biotechniques 35:862-9. 29. Hoffmaster AR, Fitzgerald CC, Ribot E, Mayer LW, Popovic T (2002) Molecular subtyping of Bacillus anthracis and the 2001 bioterrorism-associated anthrax outbreak, United States. Emerg Infect Dis 8(10):1111-6. 30. Jones SW, Dobson ME, Francesconi SC, Schoske R, Crawford R (2005) DNA assays for detection, identification, and individualization of select agent microorganisms. Croat Med J 46:522-9. 31. Chain P, Kurtz S, Ohlebusch E, Slezak T (2003) An applications-focused review of comparative genomics tools: capabilities, limitations and future challenges. Brief Bioinform 4:105-23.
Microfluidic Applications in Biodefense 367 32. Slezak T, Kuczmarski T, Ott L, Torres C, Medeiros D, Smith J, Truitt B, Mulakken N, Lam M, Vitalis E, Zemla A, Zhou CE, Gardner S (2003) Comparative genomics tools applied to bioterrorism defense. Brief Bioinform. 4:13349. 33. De BK, Bragg SL, Sanden GN, Wilson KE, Diem LA, Marston CK, Hoffmaster AR, Barnett GA, Weyant RS, Abshire TG, Ezzell JW, Popovic T (2002) A two-component direct fluorescent-antibody assay for rapid identification of Bacillus anthracis. Emerg Infect Dis 8:1060-5. 34. Belgrader P, Smith JK, Weedn VW, Northrup MA (1998) J Forensic Sciences 43:315-319. 35. Belgrader P, Benette W, Hadley D, Richards J, Stratton P, Mariella R, Milanovich F (1999) Science 284:449-450. 36. United States General Accounting Office, Report to Congressional Requesters: Bioterrorism Information Technology Strategy Could Strengthen Federal Agencies’ Abilities to Respond to Public Health Emergencies, GAO-03-139, May 2003. 37. Hindson BJ, Makarewicz AJ, Setlur US, Henderer BD, McBride MT, Dzenitis JM (2005) APDS: the autonomous pathogen detection system. Biosens Bioelectron 20:1925-31. 38. Belgrader P, Hansford D, Kovacs GT, Venkateswaran K, Mariella R Jr, Milanovich F, Nasarabadi S, Okuzumi M, Pourahmadi F, Northrup MA (1999) A minisonicator to rapidly disrupt bacterial spores for DNA analysis. Anal Chem 71:4232-6. 39. Taylor MT, Belgrader P, Furman BJ, Pourahmadi F, Kovacs GT, Northrup MA (2001) Lysing bacterial spores by sonication through a flexible interface in a microfluidic system. Anal Chem 73:492-6. 40. Belgrader P, Okuzumi M, Pourahmadi F, Borkholder DA, Northrup MA (2000) A microfluidic cartridge to prepare spores for PCR analysis. Biosens Bioelectron 14:849-52. 41. Marentis TC, Kusler B, Yaralioglu GG, Liu S, Haeggström EO, Khuri-Yakub BT(2005) Microfluidic sonicator for Q- disruption of eukaryotic cells and bacterial spores for DNA analysis. Ultrasound Med Biol 31:1265-77. 42. Hofmann O, Murray K, Wilkinson AS, Cox T, Manz A (2005) Laser induced disruption of bacterial spores on a microchip. Lab Chip 5:374-7. 43. Lee JG, Cheong KH, Huh N, Kim S, Choi JW, Ko C (2006) Microchip-based one step DNA extraction and qPCR in one chamber for rapid pathogen identification. Lab Chip 6:886-95. 44. Yin H, Killeen K (2007) The fundamental aspects and applications of Agilent HPLC-Chip. J Sep Sci 30:1427-34. 45. Kłodzińska E, Moravcova D, Jandera P, Buszewski B (2006) Monolithic continuous beds as a new generation of stationary phase for chromatographic and electro-driven separations. J Chromatogr A 1109:51-9. 46. Wu R, Hu L, Wang F, Ye M, Zou H (2007) Recent development of monolithic stationary phases with emphasis on microscale chromatographic separation. J Chromatogr A 1184:369-92.
368 Jovanovich and Horn 47. Peterson DS (2005) Solid supports for micro analytical systems. Lab Chip 5:132-9. 48. Choi S, Park JK (2005) Microfluidic system for dielectrophoretic separation based on a trapezoidal electrode array. Lab Chip 5:1161-7. 49. Tai CH, Hsiung SK, Chen CY, Tsai ML, Lee GB (2007) Automatic microfluidic platform for cell separation and nucleus collection. Biomed Microdevices 9:533-43 50. Borgatti M, Altomare L, Baruffa M, Fabbri E, Breveglieri G, Feriotto G, Manaresi N, Medoro G, Romani A, Tartagni M, Gambari R, Guerrieri R (2005) Separation of white blood cells from erythrocytes on a dielectrophoresis (DEP) based 'Lab-on-a-chip' device. Int J Mol Med 15:913-20. 51. Borgatti M, Altomare L, Abonnec M, Fabbri E, Manaresi N, Medoro G, Romani A, Tartagni M, Nastruzzi C, Di Croce S, Tosi A, Mancini I, Guerrieri R, Gambari R (2005) Dielectrophoresis-based 'Lab-on-a-chip' devices for programmable binding of microspheres to target cells. Int J Oncol 27:1559-66. 52. Wang L, Flanagan LA, Monuki E, Jeon NL, Lee AP (2007) Dielectrophoresis switching with vertical sidewall electrodes for microfluidic flow cytometry. Lab Chip 7:1114-20. 53. Yang L., Banada PP, Chatni MR, Seop Lim K, Bhunia AK, Ladisch M, Bashir RA (2006) Multifunctional micro-fluidic system for dielectrophoretic concentration coupled with immuno-capture of low numbers of Listeria monocytogenes. Lab Chip 6:896-905. 54. Wiklund M, Günther C, Lemor R, Jäger M, Fuhr G, Hertz HM (2006) Ultrasonic standing wave manipulation technology integrated into a dielectrophoretic chip. Lab Chip 6:1537-44. 55. Ryu C, Lee K, Yoo C, Seong WK, Oh HB (2003) Sensitive and rapid quantitative detection of anthrax spores isolated from soil samples by qPCR. Microbiol Immunol 47:693-9. 56. Porteous LA, Armstrong JL (1993) A simple mini-method to extract DNA directly from soil for use with polymerase chain reaction amplification. Curr Microbiol 27:115-8. 57. Kabir S, Rajendran N, Amemiya T, Itoh K (2003) Q- quantitative PCR assay on bacterial DNA: In a model soil system and environmental samples. J Gen Appl Microbiol 49:101-9. 58. Hadd, A. and S. Jovanovich. Methods and apparatus for template capture and normalization for submicroliter reaction. December 3, 2002. U. S. Patent 6,489,112. 59. Yoza B, Matsumoto M, Matsunaga T (2002) DNA extraction using modified bacterial magnetic particles in the presence of amino silane compound. J Biotechnol 94:217-24. 60. Cady NC, Stelick S, Batt CA (2003) Nucleic acid purification using microfabricated silicon structures. Biosens Bioelectron 19:59-66. 61. Christel LA, Petersen K, McMillan W, Northrup MA (1999) Rapid, automated nucleic acid probe assays using silicon microstructures for nucleic acid concentration. J Biomech Eng 121:22-7.
Microfluidic Applications in Biodefense 369 62. Liu Y, Cady NC, Batt CA (2007) A plastic microchip for nucleic acid purification. Biomed Microdevices 9:769-76. 63. Nakagawa T, Tanaka T, Niwa D, Osaka T, Takeyama H, Matsunaga T (2005) Fabrication of amino silane-coated microchip for DNA extraction from whole blood. J Biotechnol 116:105-11. 64. Chung YC, Jan MS, Lin YC, Lin JH, Cheng WC, Fan CY (2004) Microfluidic chip for high efficiency DNA extraction. Lab Chip 4:141-7. 65. Chen X, Cui D, Liu C, Li H, Chen J (2007) Continuous flow microfluidic device for cell separation, cell lysis and DNA purification. Anal Chim Acta 584:237-43. 66. Malik A (2002) Advances in sol-gel based columns for capillary electrochromatography: sol-gel open-tubular columns. Electrophoresis 23:3973-92. 67. Svec, F (2006) Less common applications of monoliths: preconcentration and solid-phase extraction. J Chromatogr B 841:52-64. 68. Breadmore MC, Wolfe KA, Arcibal IG, Leung WK, Dickson D, Giordano BC, Power ME, Ferrance JP, Feldman SH, Norris PM, Landers JP (2003) Microchip-based purification of DNA from biological samples. Anal Chem 75:1880-6. 69. Tian H, Hühmer AF, Landers JP (2000) Evaluation of silica resins for direct and efficient extraction of DNA from complex biological matrices in a miniaturized format. Anal Biochem 283:175-91. 70. Wen J, Guillo C, Ferrance JP, Landers JP (2006) DNA extraction using a tetramethyl orthosilicate-grafted photopolymerized monolithic solid phase. Anal Chem 78:1673-81. 71. Wu Q, Bienvenue JM, Hassan BJ, Kwok YC, Giordano BC, Norris PM, Landers JP, Ferrance JP (2006) Microchip-based macroporous silica sol-gel monolith for efficient isolation of DNA from clinical samples. Anal Chem 78:5704-10. 72. Wen J, Guillo C, Ferrance JP, Landers JP (2007) Microfluidic-based DNA purification in a two-stage, dual-phase microchip containing a reversed-phase and a photopolymerized monolith. Anal Chem 79:6135-42. 73. Wen J, Guillo C, Ferrance JP, Landers JP ( 2007) Microfluidic chip-based protein capture from human whole blood using octadecyl (C18) silica beads for nucleic acid analysis from large volume samples. J Chromatogr A 1171:2936. 74. Grodzinski P, Liu R, Yang J, Ward MD (2004) Microfluidic system integration in sample preparation chip-sets - a summary. In: Conf Proc IEEE Eng Med Biol Soc 4:2615-8. 75. Peoples MC, Karnes HT (in press) Microfluidic immunoaffinity separations for bioanalysis. J Chromatogr B. 76. Stevens KA, Jaykus LA (2004) Bacterial separation and concentration from complex sample matrices: a review. Crit Rev Microbiol 30:7-24. 77. O'Brien SB, Duffy G, Daly D, Sheridan JJ, Blair IS, McDowell DA (2005) Detection and recovery rates achieved using direct plate and enrichment/immunomagnetic separation methods for Escherichia coli O157:H7 in minced beef and on bovine hide. Lett Appl Microbiol 41:88-93.
370 Jovanovich and Horn 78. Fu Z, Rogelj S, Kieft TL (2005) Rapid detection of Escherichia coli O157:H7 by immunomagnetic separation and qPCR. Int J Food Microbiol 99:47-57. 79. Pyle BH, Broadaway SC, McFeters GA (1999) Sensitive detection of Escherichia coli O157:H7 in food and water by immunomagnetic separation and solid-phase laser cytometry. Appl Environ Microbiol 65:1966-72. 80. Varshney M, Yang L, Su XL, Li Y (2005) Magnetic nanoparticle-antibody conjugates for the separation of Escherichia coli O157:H7 in ground beef. J Food Prot 68:1804-11. 81. Rajkovic A, El-Moualij B, Uyttendaele M, Brolet P, Zorzi W, Heinen E, Foubert E, Debevere J (2006) Immunoquantitative qPCR for detection and quantification of Staphylococcus aureus enterotoxin B in foods. Appl Environ Microbiol 72:6593-9. 82. Ramadan Q, Samper V, Poenar D, Yu C (2006) Magnetic-based microfluidic platform for biomolecular separation. Biomed Microdevices 8:151-8. 83. Xia N, Hunt TP, Mayers BT, Alsberg E, Whitesides GM, Westervelt RM, Ingber DE (2006) Combined microfluidic-micromagnetic separation of living cells in continuous flow. Biomed Microdevices 8:299-308. 84. Furdui VI, Harrison DJ (2004) Immunomagnetic T cell capture from blood for PCR analysis using microfluidic systems. Lab Chip 4:614-8. 85. Peruski AH, Peruski LF Jr (2003) Immunological methods for detection and identification of infectious disease and biological warfare agents. Clin Diagn Lab Immunol 10:506-13. 86. Andreotti PE, Ludwig GV, Peruski AH, Tuite JJ, Morse SS, Peruski LF Jr (2003) Immunoassay of infectious agents. Biotechniques 35:850-9. 87. Fischer NO, Tarasow TM, Tok JB (2007) Heightened sense for sensing: recent advances in pathogen immunoassay sensing platforms. Analyst 132:187-91. 88. Bange A, Halsall HB, Heineman WR (2005) Microfluidic immunosensor systems. Biosens Bioelectron 20:2488-503. 89. Kim MS, Doyle MP (1992) Dipstick immunoassay to detect enterohemorrhagic Escherichia coli O157:H7 in retail ground beef. Appl Environ Microbiol 58:1764-7. 90. Grunow R, Splettstoesser W, McDonald S, Otterbein C, O'Brien T, Morgan C, Aldrich J, Hofer E, Finke EJ, Meyer H (2000) Detection of Francisella tularensis in biological specimens using a capture enzyme-linked immunosorbent assay, an immunochromatographic handheld assay, and a PCR. Clin Diagn Lab Immunol 7:86-90. 91. Herr AE, Hatch AV, Throckmorton DJ, Tran HM, Brennan JS, Giannobile WV, Singh AK (2007) Microfluidic immunoassays as rapid saliva-based clinical diagnostics. Proc Natl Acad Sci U S A 104:5268-73. 92. Schotte U, Langfeldt N, Peruski AH, Meyer H (2002) Detection of staphylococcal enterotoxin B (SEB) by enzyme-linked immunosorbent assay and by a rapid hand-held assay. Clin Lab 48:395-400. 93. Moorthy J, Mensing GA, Kim D, Mohanty S, Eddington DT, Tepp WH, Johnson EA, Beebe DJ (2004) Microfluidic tectonics platform: A colorimetric, disposable botulinum toxin enzyme-linked immunosorbent assay system. Electrophoresis 25:1705-13.
Microfluidic Applications in Biodefense 371 94. Padhye NV, Doyle MP (1991) Rapid procedure for detecting enterohemorrhagic Escherichia coli O157:H7 in food. Appl Environ Microbiol 57:2693-8. 95. Medina MB (2006) Development of a fluorescent latex microparticle immunoassay for the detection of staphylococcal enterotoxin B (SEB). J Agric Food Chem 54:4937-42. 96. Khan AS, Cao CJ, Thompson RG, Valdes JJ (2003) A simple and rapid fluorescence-based immunoassay for the detection of staphylococcal enterotoxin B. Mol Cell Probes 17:125-6. 97. Alefantis T, Grewal P, Ashton J, Khan AS, Valdes JJ, Del Vecchio VG (2004) A rapid and sensitive magnetic bead-based immunoassay for the detection of staphylococcal enterotoxin B for high-through put screening. Mol Cell Probes 18:379-82. 98. Mangru S, Bentz BL, Davis TJ, Desai N, Stabile PJ, Schmidt JJ, Millard CB, Bavari S, Kodukula K (2005) Integrated bioassays in microfluidic devices: botulinum toxin assays. J Biomol Screen 10:788-94. 99. Lim CT, Zhang Y (2007) Bead-based microfluidic immunoassays: the next generation. Biosens Bioelectron 22:1197-204. 100. Gao Y, Hu G, Lin FY, Sherman PM, Li D (2005) An electrokineticallycontrolled immunoassay for simultaneous detection of multiple microbial antigens. Biomed Microdevices 7:301-12. 101. Bai Y, Koh CG, Boreman M, Juang YJ, Tang IC, Lee LJ, Yang ST (2006) Surface modification for enhancing antibody binding on polymer-based microfluidic device for enzyme-linked immunosorbent assay. Langmuir 22:9458-67. 102. Czajka J, Batt CA (1996) A solid phase fluorescent capillary immunoassay for the detection of Escherichia coli O157:H7 in ground beef and apple cider. J Appl Bacteriol 81:601-7. 103. Peoples MC, Phillips TM, Karnes HT (2007) A capillary-based microfluidic instrument suitable for immunoaffinity chromatography. J Chromatogr B 848:200-7. 104. Herrmann M, Roy E, Veres T, Tabrizian M (2007) Microfluidic ELISA on non-passivated PDMS chip using magnetic bead transfer inside dual networks of channels. Lab Chip 7:1546-52. 105. Herrmann M, Veres T, Tabrizian M (2006) Enzymatically-generated fluorescent detection in micro-channels with internal magnetic mixing for the development of parallel microfluidic ELISA. Lab Chip 6:555-60. 106. Liu RH (2004) Integrated microfluidic biochips for immunoassay and DNA bioassays. In: Conf Proc IEEE Eng Med Biol Soc 7:5394. 107. Taitt CR, Golden JP, Shubin YS, Shriver-Lake LC, Sapsford KE, Rasooly A, Ligler FS (2004) A portable array biosensor for detecting multiple analytes in complex samples. Microb Eco 47:175-85. 108. Ligler FS, Taitt CR, Shriver-Lake LC, Sapsford KE, Shubin Y, Golden JP (2003) Array biosensor for detection of toxins. Anal Bioanal Chem 377:46977.
372 Jovanovich and Horn 109. Taitt CR, Shubin YS, Angel R, Ligler FS (2004) Detection of Salmonella enterica serovar typhimurium by using a rapid, array-based immunosensor: Appl Environ Microbiol 70:152-8. 110. McBride MT, Gammon S, Pitesky M, O'Brien TW, Smith T, Aldrich J, Langlois RG, Colston B, Venkateswaran KS (2003) Multiplexed liquid arrays for simultaneous detection of simulants of biological warfare agents. Anal Chem 75:1924-30. 111. Yang JM, Bell J, Huang Y, Tirado M, Thomas D, Forster AH, Haigis RW, Swanson, PD Wallace RB, Martinsons B, Krihak M (2002) An integrated, stacked microlaboratory for biological agent detection with DNA and immunoassays. Biosensors & Bioelectronics 17:605-618. 112. Gatto-Menking DL, Yu H, Bruno JG, Goode MT, Miller M, Zulich AW (1995) Sensitive detection of biotoxoids and bacterial spores using an immunomagnetic electrochemiluminescence sensor. Biosens Bioelectron 10:501-7. 113. Yu H, Raymonda JW, McMahon TM, Campagnari AA (2000) Detection of biological threat agents by immunomagnetic microsphere-based solid phase fluorogenic- and electro-chemiluminescence. Biosens Bioelectron 14:829-40. 114. Kijek TM, Rossi CA, Moss D, Parker RW, Henchal EA (2000) Rapid and sensitive immunomagnetic-electrochemiluminescent detection of staphyloccocal enterotoxin B. J Immunol Methods 236:9-17. 115. Rivera VR, Gamez FJ, Keener WK, White JA, Poli MA (2006) Rapid detection of Clostridium botulinum toxins A, B, E, and F in clinical samples, selected food matrices, and buffer using paramagnetic bead-based electrochemiluminescence detection. Anal Biochem 353:248-56. 116. Keener WK, Rivera VR, Young CC, Poli MA (2006) An activity-dependent assay for ricin and related RNA N-glycosidases based on electrochemiluminescence. Anal Biochem 357:200-7. 117. Smith DR, Rossi CA, Kijek TM, Henchal EA, Ludwig GV (2001) Comparison of dissociation-enhanced lanthanide fluorescent immunoassays to enzyme-linked immunosorbent assays for detection of staphylococcal enterotoxin B, Yersinia pestis-specific F1 antigen, and Venezuelan equine encephalitis virus. Clin Diagn Lab Immunol 8:1070-5. 118. Peruski AH, Johnson LH 3rd, Peruski LF Jr (2002) Rapid and sensitive detection of biological warfare agents using time-resolved fluorescence assays. J Immunol Methods 263:35-41. 119. Baselt DR, Lee GU, Natesan M, Metzger SW, Sheehan PE, Colton RJ (1998) A biosensor based on magnetoresistance technology. Biosens Bioelectron 13:731-9. 120. Millen RL, Kawaguchi T, Granger MC, Porter MD, Tondra M (2005) Giant magnetoresistive sensors and superparamagnetic nanoparticles: a chip-scale detection strategy for immunosorbent assays. Anal Chem 77:6581-7. 121. Rowe CA, Tender LM, Feldstein MJ, Golden JP, Scruggs SB, MacCraith BD, Cras JJ, Ligler FS.(1999) Array biosensor for simultaneous identification of bacterial, viral, and protein analytes. Anal Chem 71:3846-52. 122. Mulvaney SP,Cole CL, Kniller MD, Malito M, Tamanaha CR, Rife JC, Stanton MW, Whitman LJ (2007) Rapid, femtomolar bioassays in complex matri-
Microfluidic Applications in Biodefense 373 ces combining microfluidics and magnetoelectronics. Biosens Bioelectron 23:191-200. 123. Deng Y, Schwegler EE, Gravenstein S (2005) Proteomics for biodefense applications: progress and opportunities. Expert Rev Proteomics 2:203-13. 124. Chen X, Wu H, Mao C, Whitesides GM (2002) A prototype two-dimensional capillary electrophoresis system fabricated in poly(dimethylsiloxane). Anal Chem 74:1772-8. 125. Zhang C, Crasta O, Cammer S, Will R, Kenyon R, Sullivan D, Yu Q, Sun W, Jha R, Liu D, Xue T, Zhang Y, Moore M, McGarvey P, Huang H, Chen Y, Zhang J, Mazumder R, Wu C, Sobral B (2008) An emerging cyberinfrastructure for biodefense pathogen and pathogen host data. Nucleic Acids Res 36:D884-91. 126. http://www.proteomicsresource.org/ 127. Lion N, Rohner TC, Dayon L, Arnaud IL, Damoc E, Youhnovski N, Wu ZY, Roussel C, Josserand J, Jensen H, Rossier JS, Przybylski M, Girault HH. (2003) Microfluidic systems in proteomics. Electrophoresis 24:3533-62. 128. Lazar IM, Grym J, Foret F (2006) Microfabricated devices: A new sample introduction approach to mass spectrometry. Mass Spectrom Rev 25:573-94. 129. Stachowiak JC, Shugard EE, Mosier BP, Renzi RF, Caton PF, Ferko SM, Van de Vreugde JL, Yee DD, Haroldsen BL, VanderNoot VA (2007) Autonomous microfluidic sample preparation system for protein profile-based detection of aerosolized bacterial cells and spores. Anal Chem 79:5763-70. 130. Mattoon D, Michaud G, Merkel J, Schweitzer B (2005) Biomarker discovery using protein microarray technology platforms: antibody-antigen complex profiling. Expert Rev Proteomics 2:879-89. 131. Huelseweh B, Ehricht R, Marschall HJ (2006) A simple and rapid protein array based method for the simultaneous detection of biowarfare agents. Proteomics 6:2972-81. 132. Petrenko VA, Vodyanoy VJ (2003) Phage display for detection of biological threat agents. J Microbiol Methods 53:253-62. 133. Saiki RK, Scharf S, Faloona F, Mullis KB, Horn GT, Erlich HA, Arnheim N (1985) Enzymatic amplification of beta-globin genomic sequences and restriction site analysis for diagnosis of sickle-cell anemia. Science 230:1350–1354. 134. Ratcliff RM, Chang G, Kok T, Sloots TP (2007) Molecular diagnosis of medical viruses. Curr Issues Mol Biol 9:87-102. 135. Mothershed EA, Whitney AM (2006) Nucleic acid-based methods for the detection of bacterial pathogens: present and future considerations for the clinical laboratory. Clin Chim Acta 363:206-20. 136. Barken KB, Haagensen JA, Tolker-Nielsen T (2007) Advances in nucleic acid-based diagnostics of bacterial infections. Clin Chim Acta 384:1-11. 137. Ivnitski D, O'Neil DJ, Gattuso A, Schlicht R, Calidonna M, Fisher R (2003) Nucleic acid approaches for detection and identification of biological warfare and infectious disease agents. Biotechniques 35:862-9. 138. Mackay IM (2004) qPCR in the microbiology laboratory. Clin Microbiol Infect 10:190-212.
374 Jovanovich and Horn 139. Espy MJ, Uhl JR, Sloan LM, Buckwalter SP, Jones MF, Vetter EA, Yao JD, Wengenack NL, Rosenblatt JE, Cockerill FR 3rd, Smith TF (2006) qPCR in clinical microbiology: applications for routine laboratory testing. Clin Microbiol Rev 19:165-256. 140. McDonald R, Cao T, Borschel R (2001) Multiplexing for the detection of multiple biowarfare agents shows promise in the field. Mil Med 166:237-9. 141. Ramisse V, Patra G, Garrigue H, Guesdon JL, Mock M (1996) Identification and characterization of Bacillus anthracis by multiplex PCR analysis of sequences on plasmids pXO1 and pXO2 and chromosomal DNA. FEMS Microbiol Lett 145:9-16. 142. Lee MA, Brightwell G, Leslie D, Bird H, Hamilton A (1999) Fluorescent detection techniques for Q- multiplex strand specific detection of Bacillus anthracis using rapid PCR. J Appl Microbiol 87:218-23. 143. Bell CA, Uhl JR, Hadfield TL, David JC, Meyer RF, Smith TF, Cockerill FR 3rd (2002) Detection of Bacillus anthracis DNA by LightCycler PCR. J Clin Microbiol 40:2897-902. 144. Uhl JR, Bell CA, Sloan LM, Espy MJ, Smith TF, Rosenblatt JE, Cockerill FR 3rd (2002) Application of rapid-cycle Q- polymerase chain reaction for the detection of microbial pathogens: the Mayo-Roche Rapid Anthrax Test. Mayo Clin Proc 77:673-80. 145. Ryu C, Lee K, Yoo C, Seong WK, Oh HB (2003) Sensitive and rapid quantitative detection of anthrax spores isolated from soil samples by qPCR. Microbiol Immunol 47:693-9. 146. Makino S, Cheun HI (2003) Application of the qPCR for the detection of airborne microbial pathogens in reference to the anthrax spores. J Microbiol Methods 53:141-7. 147. McKillip JL, Drake M (2004) Real-time nucleic acid-based detection methods for pathogenic bacteria in food. J Food Prot 67:823-32. 148. Varma-Basil M, El-Hajj H, Marras SA, Hazbón MH, Mann JM, Connell ND, Kramer FR, Alland D (2004) Molecular beacons for multiplex detection of four bacterial bioterrorism agents. Clin Chem 50:1060-2. 149. Skottman T, Piiparinen H, Hyytiäinen H, Myllys V, Skurnik M, Nikkari S (2007) Simultaneous qPCR detection of Bacillus anthracis, Francisella tularensis and Yersinia pestis. Eur J Clin Microbiol Infect Dis 26:207-11. 150. Selvapandiyan A, Stabler K, Ansari NA, Kerby S, Riemenschneider J, Salotra P, Duncan R, Nakhasi HL (2005) A novel semiquantitative fluorescencebased multiplex polymerase chain reaction assay for rapid simultaneous detection of bacterial and parasitic pathogens from blood. J Mol Diagn 7:268-75. 151. Kim K, Seo J, Wheeler K, Park C, Kim D, Park S, Kim W, Chung SI, Leighton T (2005) Rapid genotypic detection of Bacillus anthracis and the Bacillus cereus group by multiplex qPCR melting curve analysis. FEMS Immunol Med Microbiol 43:301-10. 152. Christensen DR, Hartman LJ, Loveless BM, Frye MS, Shipley MA, Bridge DL, Richards MJ, Kaplan RS, Garrison J, Baldwin CD, Kulesh DA, Norwood DA (2006) Detection of biological threat agents by qPCR: comparison of as-
Microfluidic Applications in Biodefense 375 say performance on the R.A.P.I.D., the LightCycler, and the Smart Cycler platforms. Clin Chem 52:141-5. 153. Ulrich MP, Christensen DR, Coyne SR, Craw PD, Henchal EA, Sakai SH, Swenson D, Tholath J, Tsai J, Weir AF, Norwood DA (2006) Evaluation of the Cepheid GeneXpert system for detecting Bacillus anthracis. J Appl Microbiol 100:1011-6. 154. Zhang C, Xing D (2007) Miniaturized PCR chips for nucleic acid amplification and analysis: latest advances and future trends. Nucleic Acids Res 35:4223-37. 155. Kricka LJ, Wilding P (2003) Microchip PCR. Anal Bioanal Chem 377:820-5. 156. Wittwer CT, Fillmore GC, Hillyard DR (1989) Automated polymerase chain reaction in capillary tubes with hot air. Nucleic Acids Res 17: 4353-4357. 157. Wittwer CT, Fillmore GC, Garling DJ (1990) Minimizing the time required for DNA amplification by efficient heat transfer to small samples. Anal Biochem 186:328-31. 158. Friedman NA, Meldrum D (1998) Capillary Tube Resistive Thermal Cycling. Anal Chem 70:2997-3002. 159. Meldrum DR, Evensen HT, Pence WH, Moody SE, Cunningham DL, Wiktor PJ (2000) ACAPELLA-1K, a capillary-based submicroliter automated fluid handling system for genome analysis. Genome Res 10:95-104. 160. Northrup MA, Ching MT, White RM, Watson RT (1993) DNA amplification with a microfabricated reaction chamber. In: Digest of technical papers: transducers 1993 (Proc. 7th mt. conf. on solidstate sensors and actuators). New York Institute of Electrical and Electronic Engineers, 1993:924-6. 161. Wilding P, Shoffner MA, Kricka LJ (1994) PCR in a silicon microstructure. Clin Chem 40:1815-8. 162. Nagai H, Murakami Y, Morita Y, Yokoyama K, Tamiya E (2001) Development of a microchamber array for picoliter PCR. Anal Chem 73:1043-7. 163. Woolley AT, Hadley D, Landre P, deMello AJ, Mathies RA, Northrup MA (1998) Functional integration of PCR amplification and capillary electrophoresis in a microfabricated DNA analysis device. J Forensic Sci 43:315-9. 164. Belgrader P, Benett W, Hadley D, Richards J, Stratton P, Mariella R Jr, Milanovich F (1999) PCR detection of bacteria in seven minutes. Science 284:449-50. 165. Northrup MA, Benett B, Hadley D, Landre P, Lehew S, Richards J, Stratton P (1998) A miniature analytical instrument for nucleic acids based on micromachined silicon reaction chambers. Anal Chem 70:918-22. 166. Belgrader P, Young S, Yuan B, Primeau M, Christel LA, Pourahmadi F, Northrup MA (2000) A battery-powered notebook thermal cycler for rapid multiplex real-time PCR analysis. Biosens Bioelectron 14:849-52. 167. Neuzil P, Pipper J, Hsieh TM (2006) Disposable Q- microPCR device: labon-a-chip at a low cost. Mol Biosyst 2:292-8. 168. Neuzil P, Zhang C, Pipper J, Oh S, Zhuo L (2006) Ultra fast miniaturized qPCR: 40 cycles in less than six minutes. Nucleic Acids Res 34:e77.
376 Jovanovich and Horn 169. Wang Z, Sekulovic A, Kutter JP, Bang DD, Wolff A (2006) Towards a portable microchip system with integrated thermal control and polymer waveguides for qPCR. Electrophoresis 27:5051-8. 170. Oh KW, Park C, Namkoong K, Kim J, Ock KS, Kim S, Kim YA, Cho YK, Ko C (2005) World-to-chip microfluidic interface with built-in valves for multichamber chip-based PCR assays. Lab Chip 5:845-50. 171. Koh CG, Tan W, Zhao MQ, Ricco AJ, Fan ZH (2003) Integrating polymerase chain reaction, valving, and electrophoresis in a plastic device for bacterial detection. Anal Chem 75:4591-8. 172. Gong H, Ramalingam N, Chen L, Che J, Wang Q, Wang Y, Yang X, Yap PH, Neo CH (2006) Microfluidic handling of PCR solution and DNA amplification on a reaction chamber array biochip. Biomed Microdevices 8:167-76. 173. Liao CS, Lee GB, Wu JJ, Chang CC, Hsieh TM, Huang FC, Luo CH.(2005) Micromachined polymerase chain reaction system for multiple DNA amplification of upper respiratory tract infectious diseases. Biosens Bioelectron 20:1341-8. 174. Oda RP, Strausbauch MA, Huhmer AF, Borson N, Jurrens SR, Craighead J, Wettstein PJ, Eckloff B, Kline B, Landers JP (1998) Infrared-mediated thermocycling for ultrafast polymerase chain reaction amplification of DNA. Anal Chem 70:4361-8. 175. Roper MG, Easley CJ, Legendre LA, Humphrey JA, Landers JP.(2007) Infrared temperature control system for a completely noncontact polymerase chain reaction in microfluidic chips. Anal Chem 79:1294-300. 176. Legendre LA, Bienvenue JM, Roper MG, Ferrance JP, Landers JP (2006) A simple, valveless microfluidic sample preparation device for extraction and amplification of DNA from nanoliter-volume samples. Anal Chem 78:144451. 177. Hashimoto M, Barany F, Soper SA (2006) Polymerase chain reaction/ligase detection reaction/hybridization assays using flow-through microfluidic devices for the detection of low-abundant DNA point mutations. Biosens Bioelectron 21:1915-23. 178. Obeid PJ, Christopoulos TK, Crabtree HJ, Backhouse CJ (2003) Microfabricated device for DNA and RNA amplification by continuous-flow polymerase chain reaction and reverse transcription-polymerase chain reaction with cycle number selection. Anal Chem 75:288-95. 179. Münchow G, Dadic D, Doffing F, Hardt S, Drese KS (2005) Automated chip-based device for simple and fast nucleic acid amplification. Expert Rev Mol Diagn 5:613-20. 180. Nagai H, Murakami Y, Yokoyama K, Tamiya E.(2001) High-throughput PCR in silicon based microchamber array. Biosens Bioelectron 16:1015-9. 181. Mitterer G, Schmidt WM (2006) Microarray-based detection of bacteria by on-chip PCR. Methods Mol Biol 345:37-51. 182. Mitterer G, Huber M, Leidinger E, Kirisits C, Lubitz W, Mueller MW, Schmidt WM (2004) Microarray-based identification of bacteria in clinical samples by solid-phase PCR amplification of 23S ribosomal DNA sequences. J Clin Microbiol 42:1048-57.
Microfluidic Applications in Biodefense 377 183. Belgrader P, Benett W, Hadley D, Long G, Mariella R Jr, Milanovich F, Nasarabadi S, Nelson W, Richards J, Stratton P (1998) Rapid pathogen detection using a microchip PCR array instrument. Clin Chem 44:2191-4. 184. Northrup MA, Benett B, Hadley D, Landre P, Lehew S, Richards J, Stratton P (1998) A miniature analytical instrument for nucleic acids based on micromachined silicon reaction chambers. Anal Chem 70:918-922. 185. Higgins JA, Nasarabadi S, Karns JS, Shelton DR, Cooper M, Gbakima A, Koopman RP (2003) A handheld real time thermal cycler for bacterial pathogen detection. Biosens Bioelectron 18:1115-23. 186. McBride MT, Masquelier D, Hindson BJ, Makarewicz AJ, Brown S, Burris K, Metz T, Langlois RG, Tsang KW, Bryan R, Anderson DA, Venkateswaran KS, Milanovich FP, Colston BW Jr (2003) Autonomous detection of aerosolized Bacillus anthracis and Yersinia pestis. Anal Chem 75:5293-9. 187. Arroyo E Wheeler, EK Shediac R, Hindson B, Nasarabadi S, Vrankovich G, Bell P, Bailey C, Sheppod T, Christian AT (2005) Flow through PCR module of BioBriefcase. Smart Medical and Biomedical Sensor Technology III. Edited by Cullum BM, Chance CJ. In: Proceedings of the SPIE, Volume 6007, pp. 243-251. 188. Wilson WJ, Erler AM, Nasarabadi SL, Skowronski EW, Imbro PM. (2005) A multiplexed PCR-coupled liquid bead array for the simultaneous detection of four biothreat agents. Mol Cell Probes 19:137-44. 189. McAvin JC, McConathy MA, Rohrer AJ, Huff WB, Barnes WJ, Lohman KL (2003) A Q- fluorescence polymerase chain reaction assay for the identification of Yersinia pestis using a field-deployable thermocycler. Mil Med 168:852-5. 190. Beer NR, Hindson BJ, Wheeler EK, Hall SB, Rose KA, Kennedy IM, Colston BW (2007) On-chip, Q-, single-copy polymerase chain reaction in picoliter droplets. Anal Chem 79:8471-5. 191. Liao CS, Lee GB, Liu HS, Hsieh TM, Luo CH (2005) Miniature QPCR system for diagnosis of RNA-based viruses. Nucleic Acids Res 33:e156. 192. Marcus JS, Anderson WF, Quake SR (2006) Parallel picoliter qPCR assays using microfluidics. Anal Chem 78:956-8. 193. Deiman B, van Aarle P, Sillekens P (2002) Characteristics and applications of nucleic acid sequence-based amplification (NASBA). Mol Biotechnol 20:163-79. 194. Spargo CA, Fraiser MS, Van Cleve M, Wright DJ, Nycz CM, Spears PA, Walker GT (1996) Detection of M. tuberculosis DNA using thermophilic strand displacement amplification. Mol Cell Probes 10:247-56. 195. Hellyer TJ, Nadeau JG (2004) Strand displacement amplification: a versatile tool for molecular diagnostics. Expert Rev Mol Diagn 4:251-61. 196. Nadeau JG, Pitner JB, Linn CP, Schram JL, Dean CH, Nycz CM (1999) Anal Biochem 276:177-87. 197. Wang SS, Thornton K, Kuhn AM, Nadeau JG, Hellyer TJ (2003) Homogeneous Q- detection of single-nucleotide polymorphisms by strand displacement amplification on the BD ProbeTec ET system. Clin Chem 49:1599-607.
378 Jovanovich and Horn 198. Van Ness J, Van Ness LK, Galas DJ (2003) Isothermal reactions for the amplification of oligonucleotides. Proc Natl Acad Sci U S A 100:4504-9. 199. Tan E, Wong J, Nguyen D, Zhang Y, Erwin B, Van Ness LK, Baker SM, Galas DJ, Niemz A (2005) Isothermal DNA amplification coupled with DNA nanosphere-based colorimetric detection. Anal Chem 77:7984-92 200. Notomi T,Okayama H, Masubuchi H, Yonekawa T, Watanabe K, Amino N, Hese T (2000) Loop-mediated isothermal amplification of DNA. Nuc Acids Res 28:e63-i-vii. 201. Poon LB, Wong B, Ma, Chan K, Chow L, Abeyewickreme W, Tanngpukdee N, Yuen K, Guan Y, Looareesuwan S, Peiris J (2006) Sensitive and inexpensive molecular test for Falciparum malaria: Detecting Plasmodium falciparum DNA directly from heat-treated blood by loop-mediated isothermal amplification. Clin. Chem 52: 303-306. 202. Enosawa MS, Kageyama S, Sawai K, Watanabe K, Notomi T, Onoe S, Mori Y, Yokomizo Y (2003) Use of loop-mediated isothermal amplification of the IS900 sequence for rapid detection of cultured Mycobacterium avium subsp. Paratuberculosis. J Clin Microbiol 41:4359-4365. 203. Dean FB, Nelson JR, Giesler TL, Lasken RS (2001) Rapid amplification of plasmid and phage DNA using Phi 29 DNA polymerase and multiply-primed rolling circle amplification. Genome Res 11:1095-9. 204. Nelson JR, Cai YC, Giesler TL, Farchaus JW, Sundaram ST, Ortiz-Rivera M, Hosta LP, Hewitt PL, Mamone JA, Palaniappan C, Fuller CW (2002) TempliPhi, phi29 DNA polymerase based rolling circle amplification of templates for DNA sequencing. Biotechniques Jun;Suppl:44-7. 205. Reagin MJ, Giesler TL, Merla AL, Resetar-Gerke JM, Kapolka KM, Mamone JA (2003) TempliPhi: A sequencing template preparation procedure that eliminates overnight cultures and DNA purification. J Biomol Tech 14:143-8. 206. Hutchison CA 3rd, Smith HO, Pfannkoch C, Venter JC (2005) Cell-free cloning using phi29 DNA polymerase. Proc Natl Acad Sci U S A 102:17332-6. 207. Hashsham SA, Wick LM, Rouillard JM, Gulari E, Tiedje JM (2004) Potential of DNA microarrays for developing parallel detection tools (PDTs) for microorganisms relevant to biodefense and related research needs. Biosens Bioelectron 20:668-83. 208. Stenger DA, Andreadis JD, Vora GJ, Pancrazio JJ (2002) Potential applications of DNA microarrays in biodefense-related diagnostics. Curr Opin Biotechnol 13:208-12. 209. Volokhov D, Pomerantsev A, Kivovich V, Rasooly A, Chizhikov V (2004) Identification of Bacillus anthracis by multiprobe microarray hybridization. Diagn Microbiol Infect Dis 49:163-71. 210. Sergeev N, Distler M, Courtney S, Al-Khaldi SF, Volokhov D, Chizhikov V, Rasooly A (2004) Multipathogen oligonucleotide microarray for environmental and biodefense applications. Biosens Bioelectron 20:684-98. 211. Situma C, Hashimoto M, Soper SA (2006) Merging microfluidics with microarray-based bioassays. Biomol Eng 23:213-31. 212. Anderson RC, Su X, Bogdan GJ, Fenton J (2000) A miniature integrated device for automated multistep genetic assays. Nucleic Acids Res 28:E60.
Microfluidic Applications in Biodefense 379 213. Soper SA, Hashimoto M, Situma C, Murphy MC, McCarley RL, Cheng YW, Barany F (2005) Fabrication of DNA microarrays onto polymer substrates using UV modification protocols with integration into microfluidic platforms for the sensing of low-abundant DNA point mutations. Methods 37:103-13. 214. Liu RH, Yang J, Lenigk R, Bonanno J, Grodzinski P (2004) Self-contained, fully integrated biochip for sample preparation, polymerase chain reaction amplification, and DNA microarray detection. Anal Chem 76:1824-31. 215. Liu RH, Lodes MJ, Nguyen T, Siuda T, Slota M, Fuji HS, McShea A (2006) Validation of a fully integrated microfluidic array device for influenza A subtype identification and sequencing. Anal Chem 78:4184-93. 216. Liu RH, Nguyen T, Schwarzkopf K, Fuji HS, Petrova A, Siuda T, Peyvan K, Bizak M, Danley D, McShea A (2006) Fully integrated miniature device for automated gene expression DNA microarray processing. Anal Chem 78:19806. 217. Keen-Kim D, Grody WW, Richards CS (2006) Microelectronic array system for molecular diagnostic genotyping: Nanogen NanoChip 400 and molecular biology workstation. Expert Rev Mol Diagn 6:287-94. 218. Sosnowski R, Heller MJ, Tu E, Forster AH, Radtkey R (2002) Active microelectronic array system for DNA hybridization, genotyping and pharmacogenomic applications. Psychiatr Genet 12:181-92. 219. Borowsky J, Collins GE (2007) Chemical and biological threat-agent detection using electrophoresis-based lab-on-a-chip devices. Analyst 132:958-62. 220. Nakamura Y, Leppert M, O'Connell P, Wolff R, Holm T, Culver M, Martin C, Fujimoto E, Hoff M, Kumlin E et al (1987) Variable number of tandem repeat (VNTR) markers for human gene mapping. Science 235:1616-22. 221. Adair DM, Worsham PL, Hill KK, Klevytska AM, Jackson PJ, Friedlander AM, Keim PJ (2000) Diversity in a variable-number tandem repeat from Yersinia pestis. Clin Microbiol 38:1516-9. 222. Skuce RA, McCorry TP, McCarroll JF, Roring SM, Scott AN, Brittain D, Hughes SL, Hewinson RG, Neill SD (2002) Discrimination of M. tuberculosis complex bacteria using novel VNTR-PCR targets. Microbiology 148:519-2. 223. Keim P, LB Price AM, Klevytska, KL Smith, JM Schupp, R Okinaka, PJ Jackson, Hugh- Jones ME (2000) Multiple-locus variable-number tandem repeat analysis reveals genetic relationships within Bacillus anthracis. J Bacteriol 182:2928-36. 224. Hoffmaster AR, Fitzgerald CC, Ribot E, Mayer LW, Popovic T (2002) Molecular subtyping of Bacillus anthracis and the 2001 bioterrorism-associated anthrax outbreak, United States. Emerg. Infect Dis 8:1111-6. 225. Valjevac S, Hilaire V, Lisanti O, Ramisse F, Hernandez E, Cavallo JD, Pourcel C, Vergnaud G (2005) Comparison of minisatellite polymorphisms in the Bacillus cereus complex: a simple assay for large-scale screening and identification of strains most closely related to Bacillus anthracis. Appl Environ Microbiol 71:6613-23 226. Stratilo CW, Lewis CT, Bryden L, Mulvey MR, Bader D (2006) Singlenucleotide repeat analysis for subtyping Bacillus anthracis isolates. J Clin Microbiol 44:777-82.
380 Jovanovich and Horn 227. Vos P, Hogers R, Bleeker M, Reijans M, van de Lee T, Hornes M, Frijters A, Pot J, Peleman J, Kuiper M, et al (1995) AFLP: a new technique for DNA fingerprinting. Nucleic Acids Res 23:4407-14. 228. Janssen P, Coopman R, Huys G, Swings J, Bleeker M, Vos P, Zabeau M, Kersters K (1996) Evaluation of the DNA fingerprinting method AFLP as an new tool in bacterial taxonomy. Microbiology 142:1881-93. 229. Jackson PJ, Hill KK, Laker MT, Ticknor LO, Keim P (1999) Genetic comparison of Bacillus anthracis and its close relatives using amplified fragment length polymorphism and polymerase chain reaction analysis. J Appl Microbiol 87:263-9. 230. Lazzaro BP, Sceurman BK, Carney SL, Clark AG (2002) fRFLP and fAFLP: medium-throughput genotyping by fluorescently post-labeling restriction digestion. Biotechniques 33:539-40. 231. Bachem CW, van der Hoeven RS, de Bruijn SM, Vreugdenhil D, ZabeauM, Visser RG (1996) Visualization of differential gene expression using a novel method of RNA fingerprinting based on AFLP: analysis of gene expression during potato tuber development. Plant J 9:745-53. 232. Daffonchio D, Raddadi N, Merabishvili M, Cherif A, Carmagnola L, Brusetti L, Rizzi A, Chanishvili N, Visca P, Sharp R, Borin S (2006) Strategy for identification of Bacillus cereus and Bacillus thuringiensis strains closely related to Bacillus anthracis. Appl Environ Microbiol 72:1295-301. 233. Sacchi CT, Whitney AM, Mayer LW, Morey R, Steigerwalt A, Boras A, Weyant RS, Popovic T (2002) Sequencing of 16S rRNA gene: a rapid tool for identification of Bacillus anthracis. Emerg Infect Dis 8:1117-23. 234. Bavykin SG, Lysov YP, Zakhariev V, Kelly JJ, Jackman J, Stahl DA, Cherni A (2004) Use of 16S rRNA, 23S rRNA, and gyrB gene sequence analysis to determine phylogenetic relationships of Bacillus cereus group microorganisms. J Clin Microbiol 42:3711-30. 235. Ruppitsch W, Stöger A, Indra A, Grif K, Schabereiter-Gurtner C, Hirschl A, Allerberger F (2007) Suitability of partial 16S ribosomal RNA gene sequence analysis for the identification of dangerous bacterial pathogens. J Appl Microbiol 102:852-9. 236. Dolník V. Liu S, Jovanovich S (2000) Capillary electrophoresis on microchips (Review). Electrophoresis 21:41-54. 237. Ugaz VM, Elms RD, Lo RC, Shaikh FA, Burns MA (2004) Microfabricated electrophoresis systems for DNA sequencing and genotyping applications: current technology and future directions. Philos Transact A Math Phys Eng Sci 362:1105-29. 238. Sinville R, Soper SA (2007) High resolution DNA separations using microchip electrophoresis. J Sep Sci 30:1714-28. 239. Chen YH, Wang WC, Young KC, Chang TT, Chen SH (1999) Plastic microchip electrophoresis for analysis of PCR products of hepatitis C virus. Clin Chem 45:1938-43. 240. Harrison DJ, Fluri K, Seiler K, Fan Z, Effenhauser CS, Manz A (1993) Micromachining a Miniaturized Capillary Electrophoresis-Based Chemical Analysis System on a Chip. Science 261:895-897.
Microfluidic Applications in Biodefense 381 241. Woolley AT, Mathies RA (1994) Ultra-high-speed DNA fragment separations using microfabricated capillary array electrophoresis chips. Proc Natl Acad Sci USA 91:11348-11352. 242. Jacobson SC, Hergenroder R, Koutny LB, Ramsey JM (1994) High-speed separations on a microchip. Anal Chem 66:1114-1118. 243. Woolley AT, Sensabaugh GF, Mathies RA (1997) High-Speed DNA Genotyping Using Microfabricated Capillary Array Electrophoresis Chips. Anal Chem 69:2181-2186. 244. Effenhauser CS, Paulus A, Manz A, Widmer HM (1994) High-Speed Separation of Antisense Oligonucleotides on a Micromachined Capillary Electrophoresis Device. Anal. Chem. 66:2949-2953. 245. Liu SR, Shi YN, Ja WW, Mathies RA (1999) Optimization of high-speed DNA sequencing on microfabricated capillary electrophoresis channels. Anal. Chem 71:566-573. 246. Shen Z, Liu X, Long Z, Liu D, Ye N, Qin J, Dai Z, Lin B (2006) Parallel analysis of biomolecules on a microfabricated capillary array chip. Electrophoresis 27:1084-92. 247. Liu S, Ren H, Gao Q, Roach DJ, Loder Jr RT, Armstrong TM, Mao Q, Blaga I, Barker DL, Jovanovich SB (2000) Automated parallel DNA sequencing on multiple channel chips. Proc Natl Acad Sci USA 97: 5369-5374. 248. Kartalov EP, Quake SR (2004) Microfluidic device reads up to four consecutive base pairs in DNA sequencing-by-synthesis. Nucleic Acids Res 32:28739. 249. Russom A, Tooke N, Andersson H, Stemme G (2005) Pyrosequencing in a microfluidic flow-through device. Anal Chem 77:7505-11. 250. Seiler K, Harrison D, Manz A (1993) Micromachining a Miniaturized Capillary Electrophoresis-Based Chemical Analysis System on a Chip. Science 261:895-897. 251. Vreeland WN, Williams SJ, Barron AE, Sassi AP (2003) Tandem isotachophoresis-zone electrophoresis via base-mediated destacking for increased detection sensitivity in microfluidic systems. Anal Chem 75:3059-65. 252. Wainright A, Nguyen UT, Bjornson T, Boone TD (2003) Preconcentration and separation of double-stranded DNA fragments by electrophoresis in plastic microfluidic devices. Electrophoresis 24:3784-92. 253. Jung B, Bharadwaj R, Santiago JG (2006) On-chip millionfold sample stacking using transient isotachophoresis. Anal Chem 78:2319-27. 254. Lichtenberg J, Verpoorte E, de Rooij NF (2001) Sample preconcentration by field amplification stacking for microchip-based capillary electrophoresis. Electrophoresis 22:258-71. 255. Gong M, Wehmeyer KR, Limbach PA, Arias F, Heineman WR (2006) Online sample preconcentration using field-amplified stacking injection in microchip capillary electrophoresis. Anal Chem 78:3730-7. 256. Simpson PC, Roach D, Woolley AT, Thorsen T, Johnston R, Sensabaugh GF, Mathies RA (1998) High-throughput genetic analysis using microfabricated 96-sample capillary array electrophoresis microplates. Proc Natl Acad Sci U S A 95:2256-61.
382 Jovanovich and Horn 257. Cheng J, Waters LC, Fortina P, Hvichia G, Jacobson SC, Ramsey JM, Kricka LJ, Wilding P (1998) Degenerate oligonucleotide primed-polymerase chain reaction and capillary electrophoretic analysis of human DNA on microchipbased devices. Anal Biochem 257:101-106. 258. Schmalzing D, Koutny L, Adourian A, Belgrader P, Matsudaira P, Ehrlich D (1997) DNA typing in thirty seconds with a microfabricated device. Proc Natl Acad Sci USA 94:10273-10278. 259. Emrich CA, Tian H, Medintz IL, Mathies RA (2002) Microfabricated 384lane capillary array electrophoresis bioanalyzer for ultrahigh-throughput genetic analysis. Anal Chem 74:5076-83. 260. Woolley AT, Mathies RA (1994) Ultra-high-speed DNA fragment separations using microfabricated capillary array electrophoresis chips. Proc Natl Acad Sci U S A 91:11348-52. 261. Medintz IL, Paegel BM, Mathies RA (2001) Microfabricated capillary array electrophoresis DNA analysis systems. J Chromatogr A 924:265-70. 262. Paegel BM, Emrich CA, Wedemayer GJ, Scherer JR, Mathies RA (2002) High throughput DNA sequencing with a microfabricated 96-lane capillary array electrophoresis bioprocessor. Proc Natl Acad Sci U S A 99:574-9. 263. Salas-Solano O, Schmalzing D, Koutny L, Buonocore S, Adourian A, Matsudaira P, Ehrlich D (2000) Optimization of high-performance DNA sequencing on short microfabricated electrophoretic devices. Anal Chem 72:3129-37. 264. Lin YW, Chang HT (2005) Modification of poly(methyl methacrylate) microchannels for highly efficient and reproducible electrophoretic separations of double-stranded DNA. J Chromatogr A.1073:191-9. 265. Dang F, Tabata O, Kurokawa M, Ewis AA, Zhang L, Yamaoka Y, Shinohara S, Shinohara Y, Ishikawa M, Baba Y (2005) High-performance genetic analysis on microfabricated capillary array electrophoresis plastic chips fabricated by injection molding. Anal Chem 77:2140-6. 266. Zhou XM, Dai ZP, Liu X, Luo Y, Wang H, Lin BC (2005) Modification of a poly(methyl methacrylate) injection-molded microchip and its use for high performance analysis of DNA. J Sep Sci 28:225-33. 267. Ricco AJ, Boone TD, Fan ZH, Gibbons I, Matray T, Singh S, Tan H, Tian T, Williams SJ (2002) Application of disposable plastic microfluidic device arrays with customized chemistries to multiplexed biochemical assays. Biochem Soc Trans 30:73-8. 268. Llopis SD, Stryjewski W, Soper SA (2004) Near-infrared time-resolved fluorescence lifetime determinations in poly(methylmethacrylate) microchip electrophoresis devices. Electrophoresis 25:3810-9. 269. Pumera M (2006) Analysis of nerve agents using capillary electrophoresis and laboratory-on-a-chip technology. J Chromatogr A 1113:5-13. 270. Wang J, Zima J, Lawrence NS, Chatrathi MP, Mulchandani A, Collins GE (2004 ) Microchip capillary electrophoresis with electrochemical detection of thiol-containing degradation products of V-type nerve agents. Anal Chem 76:4721-6.
Microfluidic Applications in Biodefense 383 271. Fruetel JA, Renzi RF, Vandernoot VA, Stamps J, Horn BA, West JA, Ferko S, Crocker R, Bailey CG, Arnold D, Wiedenman B, Choi WY, Yee D, Shokair I, Hasselbrink E, Paul P, Rakestraw D, Padgen D (2005) Microchip separations of protein biotoxins using an integrated hand-held device. Electrophoresis 26:1144-54. 272. Renzi RF, Stamps J, Horn BA, Ferko S, Vandernoot VA, West JA, Crocker R, Wiedenman B, Yee D, Fruetel JA (2005) Hand-held microanalytical instrument for chip-based electrophoretic separations of proteins. Anal Chem 77:435-4. 273. Woolley AT, Hadley D, Landre P, deMello AJ, Mathies RA, Northrup MA (1996) Functional integration of PCR amplification and capillary electrophoresis in a microfabricated DNA analysis device. Anal Chem 68:4081-6. 274. Swerdlow H, Jones BJ, Wittwer CT (1997) Fully automated DNA reaction and analysis in a fluidic capillary instrument. Anal Chem 69:848-55. 275. He Y, Zhang YH, Yeung ES (2001) Capillary-based fully integrated and automated system for nanoliter polymerase chain reaction analysis directly from cheek cells. J Chromatogr A 924:271-84. 276. Zhang Y, He Y, Yeung ES (2001) High-throughput polymerase chain reaction analysis of clinical samples by capillary electrophoresis with UV detection. Electrophoresis 22:2296-302. 277. Zhang N, Yeung ES (1998) On-line coupling of polymerase chain reaction and capillary electrophoresis for automatic DNA typing and HIV-1 diagnosis. J Chromatogr B 714:3-11. 278 Hashimoto M, He Y, Yeung ES (2003) On-line integration of PCR and cycle sequencing in capillaries: from human genomic DNA directly to called bases. Nucleic Acids Res 31:e41. 279. Burns MA, Johnson BN, Brahmasandra SN, Handique K, Webster JR, Krishnan M, Sammarco TS, Man PM, Jones D, Heldsinger D, Mastrangelo CH, Burke DT (1998) An integrated nanoliter DNA analysis device. Science 282:484-7. 280. Lagally ET, Emrich CA, Mathies RA (2001) Fully integrated PCR-capillary electrophoresis microsystem for DNA analysis. Lab Chip 1:102-7. 281. Lagally ET, Scherer JR, Blazej RG, Toriello NM, Diep BA, Ramchandani M, Sensabaugh GF, Riley LW, Mathies RA (2004) Integrated portable genetic analysis microsystem for pathogen/infectious disease detection. Anal Chem 76:3162-70. 282. Lagally ET, Medintz I, Mathies RA (2001) Single-molecule DNA amplification and analysis in an integrated microfluidic device. Analytical Chemistry 73:565-570. 283. Blazej R, Kumaresan P, Mathies RA (2006) Microfabricated bioprocessor for integrated nanoliter-scale Sanger DNA sequencing. PNAS 103: 7240–45. 284. Easley CJ, Karlinsey JM, Landers JP (2006) On-chip pressure injection for integration of infrared-mediated DNA amplification with electrophoretic separation. Lab Chip 6:601-10. 285. Easley CJ, Karlinsey JM, Bienvenue JM, Legendre LA, Roper MG, Feldman SH, Hughes MA, Hewlett EL, Merkel TJ, Ferrance JP, Landers JP (2006) A
384 Jovanovich and Horn fully integrated microfluidic genetic analysis system with sample-in-answerout capability. Proc Natl Acad Sci U S A 103:19272-7. 286. (2006) “Nucleic Acid-Based Technologies Ramp Up Alternative Methods for Amplifying and Detecting Specific DNA Sequences,” Genetic Engineering News. 26(12). 287. Chan EY, Goncalves NM, Haeusler RA, Hatch AJ, Larson JW, Maletta AM, Yantz GR, Carstea ED, Fuchs M, Wong GG, Gullans SR, Gilmanshin R (2004) DNA mapping using microfluidic stretching and single-molecule detection of fluorescent site-specific tags. Genome Res 14:1137-46. 288. Grover WH, Mathies RA (2005) An integrated microfluidic processor for single nucleotide polymorphism-based DNA computing. Lab Chip 5:1033-40. 289. Ohman O, Jacobson G, Ekstrom B (1993) Miniaturized electrophoresis in planar foramtas produced by semiconductor fabrication techniques. In: Program and abstracts ICES-ELPHO 93. Sandefjord, Norway, June 2-4, 1993:24.
Chapter 11 Current and Future Trends in Microfluidics within Biotechnology Research
Abraham P. Lee1,2 and Gisela Lin2 University of California at Irvine, Biomedical Engineering Department, Irvine, CA 92697 1
2
Micro/nano Fluidics Fundamental Focus (MF3) Center
Correspondence should be addressed to: Abraham Lee (
[email protected])
Keywords: Microfluidics, biotechnology, bioprocessing, pumps, valves, genomics, microarrays, cell biology, microchannel, microTAS, biomedicine, diagnostics
Abstract Microfluidics is a field that has been growing ever since the early 1990s. Microfluidic devices originated from the integrated circuits (IC) industry, and early microfluidic devices are mostly silicon based analyzers consisting of channels for species separation in a carrier fluid. However, with recent advances in microfabrication and device design innovative components and platforms are coming to fore with applications in biology, medicine, pharmaceutics, and food and environmental monitoring. “Micro total analysis systems” (microTAS) or “labs-on-a-chip” (LOC) are becom-
386
Lee and Lin
ing a reality where entire chemical analyses in miniaturized volumes are performed with high sensitivities and in shorter time spans. This chapter explores the future directions this field is taking. New materials, fabrication techniques, and device designs are streamlining the way biological experiments are done in life science laboratories as well as in nonlaboratory settings. Microfluidic devices are becoming more integrated, providing new capabilities at the microscale and facilitating the commercialization of these platforms. Much like how the integrated circuits (IC) chips revolutionized the way we live and communicate, a total integrated fluidic chip would likely create applications and functions that we cannot begin to imagine today.
11.1 The Past – Exciting Prospects Andreas Manz introduced the vision of a “micro total analysis system” (micro TAS) in 1990 [1] where on-chip components would be able to carry out chemical analyses in miniaturized volumes with high sensitivities and in shorter times. Since then, the field of microfluidics has never stopped growing with innovative components and platforms ever expanding into more applications in biology, medicine, pharmaceutics, and food and environmental monitoring. The field has also benefited tremendously from the participation of researchers from a broad spectrum of disciplines, ranging from chemists, biologists, material scientists, physicists, chemical engineers, mechanical engineers, electrical engineers, and biomedical engineers. Numerous large corporations have set up R&D divisions to explore commercialization opportunities in microfluidics. Even large companies traditionally outside of the biotechnology field have jumped on the bandwagon. For example, Intel Corporation launched a Precision Biology group. Others (e.g. GE, Honeywell, Samsung etc.) have initiated research projects to explore the commercialization potential of microfluidics. After an initial explosion of start-up companies during the dot-com era (1998-2001) and the subsequent bubble-bursting period (2001-2005), the number of start-up companies based on microfluidics has started to recover and has steadily increased up to the present (2006-2008). No fewer than 20 startup companies in the US are either focusing on microfluidics enabled products or incorporating microfluidic technologies to produce/manufacture their end product.
Current and Future Trends
387
In the midst of all these exciting developments, however, there have been very few “microTAS” that actually entail entire processes from sample input to sample preparation to sample detection. Mostly components for singular functions have been developed (e.g. pumps, separation, or detection, etc.). In many practical applications (e.g. drug discovery and screening), a total analysis system is not necessary to achieve the intended functions. However, much like how the integrated circuits (IC) chips revolutionized the way we live and communicate, a total integrated fluidic chip would likely create applications and functions that we cannot begin to imagine today [2].
a)
b)
Fig. 11.1 Fabrication of silicon microfluidic devices. (a) Cross channel design of the first CE on chip in silicon fabricated by wet etching (From [9]. Reprinted with permission from AAAS) . (b) DRIE of silicon maze microchannel network for biocomputation study Reprinted with permission from [52].
Microfluidic devices in their current form originated from the integrated circuits (IC) industry, and the early microfluidic devices towards µTAS are mostly silicon based analyzers consisting of channels for species separation in a carrier fluid. Photolithography techniques make it easy to batch fabricate complicated 2-D microfluidic networks with just change of the mask design. Figure 11.1a shows the first microfluidic on-chip cross channels for capillary electrophoresis (CE), fabricated by wet etching of silicon. A more complicated maze fluidic network for a biocomputation study, fabricated by silicon deep reactive ion etching (DRIE), is shown in Fig. 11.1b. With advancements in microfabrication technology, many other techniques such as soft lithography [3], LIGA techniques [4], laser micromachining [5], electron beam lithography [6], and dip pen nanolithography [7] have broadened the field and made it possible to fabricate diverse micro- and nano- devices to meet the need of chemical and biological applications. Polymer materials become more and more attractive because they are low cost and easy to fabricate. Replication, inject molding, hot embossing, stereolithography, and layering techniques have been
388
Lee and Lin
demonstrated to fabricate microfluidic devices from various polymer materials [8]. With the invention of the soft lithography technique [3], fast prototyping of microfluidic devices can be demonstrated in polydimethylsiloxane (PDMS) fluidic channels that minimize complicated cleanroom fabrication steps. Most importantly, large scale integration in microfluidics is enabled by utilizing the flexible mechanical property of PDMS for multilayer structures with fluidic pumps and valves fabricated in PDMS termed the multiple layer soft lithography (MSL) technique. PDMS based microfluidic devices have become the most widely used method in microfluidics research labs for proof-of-concept demonstrations. PDMS is even being developed for some industrial applications [10]. Unlike in electronics, where the electrons or holes are transferred in the electric circuits, different reagents, buffers, solvents, and solutions are the objects of interest to be transported in microchannels for microfluidic applications. Due to the complex nature of the solutions faced in microfluidics, more design considerations are required to assure that the channel materials are compatible with the sample solutions. Although PDMS has showed great success in various applications of microfluidics, new materials and fabrication methods have to be found. Recently, a solvent resistant material PFPE has been used to replace PDMS for harsh solvent reaction and it not only provides the advantage of PDMS with excellent mold replication ability and transparency to light, but it is also resistance to most of the solvents that attack PDMS [11]. Since different microfluidic applications often require different materials and surface properties, they are generally more complicated than for IC applications where silicon is more or less considered a universal material. New materials and new fabrication techniques have to be explored to develop microfluidic systems for various biotechnology applications.
11.2 The Present – Kaleidoscope-like Trends The current state of microfluidics is much like a kaleidoscope, appealing to the eye and radiating in different directions, yet lacking a firm track that will focus all the beaming efforts in a meaningful and lucrative fashion. The technology has foundations similar to those of electronics and photonics, but the challenge is much greater since the applications are more diverse and the constituents have very different physical and chemical prop-
Current and Future Trends
389
erties. Research efforts at academic institutions have been prolific in studying microfluidic phenomena and producing a variety of novel microfluidic devices. The microfabrication of these devices is predominantly in the following categories, soft lithography of PDMS, photolithography and silicon/glass micromachining, and other polymer molding techniques (e.g. hot embossing or injection molding). Most of the devices focus primarily on the structures of microchannels, reservoirs, and membranes. Active components and energy sources such as syringe pumps, pneumatic pressure sources, and mechanical plungers are mostly connected from off-chip. Most recently, there have been some efforts to truly integrate the active components on-chip. For example, the Whitesides group developed various techniques to integrate (co-fabricate) optical and electrical components on-chip [12, 13]. The Lee lab and others have developed side-wall vertical electrodes for lateral deflection of particulates in microchannels as well as magnetohydrodynamic (MHD) pumping [14]. Chang Liu’s group developed a platform that interconnects the passive and active fluidic control elements in a breadboard-like platform [15]. The following are current research areas that we believe will make an impact going forward into the future. 11.2.1 Droplet microfluidics Conventional microfluidics, or “analog” microfluidics, is governed by diffusion and low Reynolds number laminar flows that prevent rapid mixing of miscible liquids. In comparison, droplet “digital” microfluidics discretizes the fluids into femtoliter to nanoliter volumes separated by immiscible fluids. The separation involves the disruption of interfacial forces between the two fluids either via flow induced shear forces or other surface forces, such as electrowetting or dielectrophoresis (DEP). The sizes of the droplets are extremely monodisperse (within 2%) and can be controlled over a wide range (typically 1µm to 100µm in diameter). Several review papers cover the details of this exciting technology very well [16-20]. This method produces highly monodisperse droplets in the nanometer to micrometer diameter range, and particles from a variety of materials at rates over twenty thousand per second [21]. Due to high surface area to volume ratios at the microscale, heat and mass transfer times and diffusion distances are shorter, facilitating faster reaction times. Droplet-based microfluidics allows for independent control of each droplet, thus generating microreactors that can be individually transported, mixed, and analyzed [22]. Since multiple identical microreactor units can be formed in a short time,
390
Lee and Lin
parallel processing and experimentation can easily be achieved, allowing large data sets to be taken efficiently and enabling high throughput screening and synthesis. Droplet microfluidics also offers greater potential for increased throughput and scalability than other continuous-flow approaches. In the past 5 years, several groups have used digital microfluidics to form irregular particles [23], double emulsions [24], hollow microcapsules [25], and microbubbles [26]. The current major applications of digital microfluidics lie in two general directions, droplets for cellular and molecular assays and droplets for generating micro/nanoparticles for advanced therapeutics. The former application enables the rapid screening of cells [27-29], protein crystallization conditions [30-33], nucleic acids [34, 35], and PCR in droplets [35, 36]. These platforms can be generated on demand and can be programmable. Thousands of different conditions and combinations can be screened to find optimized ones for genomics, proteomics, and cellomics. The latter application utilizes the uniform and rapid mixing of precursor reagents inside droplets to produce monodispersed particles and vesicles. Control over particle size and minimization of the size distribution is important for the use of particles as a route of administration and controlled release of encapsulated materials such as drugs, dyes, or enzymes, etc [37]. Depending on what reagents are used, the droplets can either be solidified (e.g. polymerize) into solid particles or be annealed leading to smaller particles precipitating inside the droplets. Microfluidic devices can also produce multilayer vesicles where a polymer or lipid membrane forms around an encapsulated gas, liquid, or solid core. The aforementioned particles or vesicles are possible due to the intrinsic stability of laminar flow microfluidics and the ability to “nanomanufacture” various droplet sizes and form factors. One of the challenges is to generate the particles/vesicles in sufficiently large quantities for biotechnology and biomedical applications. Several groups have generated quantum dots using microfluidic droplets for better control of the reaction parameters [38, 39]. Biodegradable particles can be used for controlled release of drugs and biologics (including poly(lactic-co-glycolic acid) (PLGA) nanoparticles for protein drug delivery [40] and calcium alginate beads through controlled droplet fusion [41]). Others have envisioned using these particles for the delivery of stem cells for regenerative medicine [42]. Hydrogels such as alginate and agarose [43, 44] are the most commonly used polymers to carry therapeutic stem cells because of their ease in gel formation. Another recent development is the utilization of microfluidic technologies for the generation of gas encapsulating microbubbles with lipid shells for ultrasound contrast
Current and Future Trends
391
imaging agents [45]. Microbubbles have been used as ultrasound contrast agent in ultrasonic imaging to improve the accuracy of cancer detection [46, 47]. They have a proven clinical utility, particularly as a diagnostic tool in cardiology [48], radiology [49], and oncology [50]. The gas composition and membrane components determine the stability of the microbubble contrast agents and how long the microbubbles can be sustained in the body. The importance and success of using microfluidic systems for these biotechnological applications stem from their excellent ability to control size and composition uniformity. 11.2.2 Integrating Active Components in Microfluidics As mentioned earlier, the field of microfluidics is predominantly focused on passive microfluidic channel structures without the integration of active energy components. So far, most of the active microfluidic components were typically fabricated on silicon and glass using processing techniques borrowed from the IC industry. The processes were not as suitable for many biotechnological applications due to the cost and time required for the fabrication as well as the incompatibility with biological constituents. The recent surge of polymeric devices for microfluidics leads to the natural search for microfabrication techniques that incorporate electrical and optical components on polymeric microfluidic channels. A previously published book chapter reported on some of the active “microelectrofluidic” devices [51]. Figures 11.2 and 11.3 [52] illustrate a vision of the complete bioprocessing station that is implemented by electrodes-on-fluidic chips. Critical issues to deal with include: the electrochemical integrity of the electrodes in fluids, the immobilization of biomolecules on electrodes, the ability to produce vertical electrodes to deflect particles and molecules laterally to different channels, alignment of electrodes to the channels, and the robustness of electrical interconnects. As illustrated, the power of microelectrofluidics lies in its ability to process and prepare samples for multiple analysis steps with a single integrated platform.
392
Lee and Lin
Micro electrofluidic network Pump
electroporation
Pump Pump
Recirculation separator
DEP separator MHD switches MHD
mixing heater
sensing
PCR
DEP trapping
Pump
Fig. 11.2 Schematic illustrating the various components that can be implemented by a micro electrofluidic network.
Fig. 11.3 Microfluidic branched channels with electrodes embedded.
As microfluidic devices are generally two dimensional layouts resulting in channels splitting and merging on a planar substrate, it is desirable to be able to manipulate fluids and their constituents in the 2-D plane. This means moving the fluids/constituents laterally to direct them into different channels. The miniaturization of electrodes in microfluidic devices scales favorably for electrical forces (e.g. DEP, electrokinetic, or electrostatic) as they require less power and microfabrication techniques enable their integration with various lab-on-a-chip components. Most micro electrofluidic configurations utilize planar electrodes on the floor of the microchannel to generate electrical field forces along the vertical direction, such as those used in field-flow-fractionation (FFF)-DEP for separation and sorting of
Current and Future Trends
393
cells [53, 54]. However, for thin film planar electrodes there is always dead electrical field space at a distance away from the bottom of the microchannel and the electrical (e.g. DEP or electrostatic) force becomes less dominant above the middle height of the channel. Fiedler et al. utilized planar electrodes on the bottom and top layer of the microchannel to achieve aligning, caging and switching [55]. High aspect ratio electrodes extending along the height of the channel can also be implemented to extend the electrical fields to the whole channel space, avoiding dead electrical field space. For this purpose, different 3-D electrodes have been developed and demonstrated. Iliescu et al. [56] reported highly doped 3-D silicon electrodes. Voldman et al. developed electroplated 3D metal pillar DEP devices for cell trapping [57] and construction of a dynamic array cytometer [58]. Madou has developed 3-D carbon SU-8 electrodes for DEP extraction of nanofibrous carbon from oil [59]. Holmes et al. put three pairs of electrodes at the junction of a microchannel to achieve two-way sorting [60]. Our group has developed a microfabrication process for vertical electrodes in the sidewalls of microchannels to generate lateral DEP forces in the channel. With the electrodes in the sidewalls and extending along the entire height of the channel, a relatively uniform electrical field gradient along the width of the channel can be generated. The microfabrication process was utilized to demonstrate MHD [61] and DEP [62] switching of particles to different outlet channels. For the DEP switching, the same frequency but different amplitude voltages were applied to two opposing sets of interdigitated electrodes to switch particles laterally. The Whitesides group has developed a process termed “microsolidics” that utilizes PDMS channels filled with solder to generate 3-D electrodes that can be integrated with microfluidic channels [13]. They have demonstrated microcoils for heaters in microchannels as well as magnets [12] in microfluidic channels where they demonstrated capture/release and sorting of superparamagnetic particles. These structures make it possible to produce electrical fields (with strength approaching 108 V•m-1) across microfluidic channels by applying an electrical potential between two wires separated by a channel. Techniques have been developed [63, 64] to generate low voltage, elastomer-metal ‘wet’ actuators that take the place of the pneumatic actuators of Steve Quake’s PDMS valves [65]. Most recently, the Whitesides group has been making optical devices in microfluidics. By taking advantage of the laminar flow of fluids at the microscale, fluids with different refractive index can be flowed side by side in a microchannel without mixing. Thus, fluidic waveguides can be created by flowing an “optical core” fluid flanked by two “optical cladding” fluids [66, 67]. The liquid-liquid waveguide is formed from a single step of soft
394
Lee and Lin
lithography, and thus can be easily integrated with the electrical, magnetic, and fluidic components previously described. In addition, a key advantage of a liquid-liquid waveguide is that the optical interfaces forming the waveguide are the smooth boundaries defined by liquids under laminar flow. Accordingly, the requirements for the roughness of the channels are substantially lower than those for conventional waveguides. The waveguide technology has been extended to fabricate microfluidic dye lasers [68]. The laser diagram is shown in Fig. 11.4. The properties of the laser can be easily adjusted by changing the refractive index contrast between core and cladding fluids.
Fig. 11.4 Top-view scheme for an L2 waveguide laser consisting of a liquid cladding and fluorescent liquid core flowing laminarly in a microfluidic channel in PDMS. The dimensions of the central channel were 100µm × 400µm × 10 mm. (height × width × length). The 532-nm laser beam (frequency doubled Nd:YAG, 50 Hz repetition rate, 16-ns pulse) was elongated with a cylindrical lens, and the optical pumping region covered the full length of the waveguiding region for a 10mm-long channel. (Reprinted with permission from [66]. Copyright 2005 American Chemical Society).
11.2.3 Third world - paper microfluidics – George Whitesides In terms of ultra-low cost, low-tech microfluidic platforms, the Whitesides group has been exploring the use of Ppaper for microfluidics [69]. By simply patterning chromatography paper using Su-8 photoresist, channels as small as 100um in width have been defined. Since the paper is hydrophilic and the photoresist is hydrophobic, liquids are easily confined to these pa-
Current and Future Trends
395
per channels. No power sources are required as the samples “wick” into the channels travel towards assay reservoirs where reagents and other analysis chemicals are applied. The result of the assay is a color change in the reservoir, which can be detected optically using a microscope, comparison to a colormetric strip, or by using a simple optical scanner. Paper microfluidic platforms are easily handled, transported, and analyzed. This type of patterned paper could become the basis for low-cost, portable, and technically simple multiplexed bioassays. The Whitesides laboratory has demonstrated its capability by the simultaneous detection of glucose and protein in 5 mL of urine. The assay system is small, disposable, easy to use (and carry), and requires no external equipment, reagents, or power sources. This platform is attractive for use in less industrialized countries, in the field, or as an inexpensive alternative to more-advanced technologies already used in clinical settings. 11.2.4 Microfluidic solutions for enhancing existing biotechnology platforms Large scale DNA analysis was made possible by DNA microarrays that were first commercialized by Affymetrix [70]. A parallel development was the capillary electrophoresis devices that enabled separation of DNA fragments for the assaying of biological material. Biomolecular (DNA, RNA, or protein) microarrays have become very powerful tools for biologists, medical researchers, and even clinical practitioners. They have been useful for studying mRNA levels, genetic mutations, genotyping, and examining gene expression in biological samples. They are also slowly but surely being used at hospitals and medical labs as a powerful tool in the arena of diagnostics and personalized medicine [71]. However, most of the microarrays use platforms that rely on a desktop machine with robotics to dispense the liquids and rely on diffusion for target molecules to hybridize with probe molecules immobilized on the arrays. The integration with microfluidics will greatly empower these platforms to be automated, with higher throughput and on smaller footprints. These integrated microfluidic systems will enable many distributed detection systems such as point-of-care or fieldable diagnostics that can be used in remote (e.g. third world health) and extreme (e.g. battlefield or space aircrafts) environments. Current efforts include the use of microfluidics for pharmacogenomics and molecular diagnostics applications. For the former effort, genetic screening is used to decipher the effectiveness of drugs on certain diseases by analyzing the regulated genes and detecting critical polymorphisms or mutations in a
396
Lee and Lin
gene sequence. The categories of the latter effort include genetic screening for disease susceptibility and the detection of pathogens in human or environmental samples. Very few groups have reported on integrating DNA microarrays with microfluidics. One such system was designed by Robin Liu’s group [71-73]. The Combi-Matrix microarray chip (CustomArrayTM) was coupled to a microfluidic plastic cartridge to form a microfluidic microarray device. Active semiconductor circuit elements in the design permit the selection and parallel activation of individual electrodes in the array to perform in situ oligonucleotide synthesis of customized content on the chip. The oligonucleotides were synthesized by using these electrodes to electrochemically generate protons to remove the blocking chemistry cap for selective reaction with the nucleotide reagents. This method resulted in a fully integrated and self-contained microfluidic biochip device to automate the fluidic handling steps required to carry out microarray-based gene expression or genotyping analysis. Many of these integrated systems have the potential to be applied to protein microarrays as well [74-76]. There have been several novel methods developed to sort and label individual cells and molecules, which allows the creation and manipulation of surfaces and tags at the cellular/molecular level. For example, Keating and co-workers have produced nanowire “bar codes” made of customized patterns of metallic bands that are attached to specific molecules. These nanowires are made by electrodepositing metals into the pores of alumina template membranes. The surface of the bar-coded nanowire is coated with probes, and each barcode is read optically. Solution arrays of highly encodable and chemically diverse nanowires offer an interesting solution to the problems of carrying out comprehensive analyses on classes of biologically relevant molecules (e.g., genomics, proteomics, etc.) and the demand for very high-level multiplexing in small sample volumes. Combined with microfluidics, bar-coded nanowires could be a powerful tool for high-throughput molecular detection and sorting [77, 78] Dr. Mehmet Toner at Harvard University has taken an alternative approach to encoding particles for high-throughput biomolecule analysis. Poly(ethylene glycol) (PEG) particles are simultaneously created and encoded as the laminar flow of multiple streams containing molecular probes and fluorescent markers pass over the front surface of the particles. Meanwhile UV lithography occurs through the backside, polymerizing the polymer into flat, “pill-shaped” particles. Multiple probes can be placed on each particle by flowing additional adjacent streams of different mole-
Current and Future Trends
397
cules. This method can generate multifunctional particles bearing over a million unique codes. Once the particles are created, they flow downstream through a detection microchannel, where the particles are aligned and scanned optically. Applications of these microparticles include genetic analysis, combinatorial chemistry, and clinical diagnostics.[79] Microfluidics has also proved beneficial in chemical analysis systems such as liquid chromatography (LC). For example, Woolley and coworkers demonstrated rapid microchip LC analysis with integrated electrolysisbased pumping. Three glass microscope slides were sandwiched together to make the integrated device. The fabrication process employed a combination of photolithography, wet chemical etching, through-drilling with diamond-tipped bits, and thermal bonding. Laser-induced fluorescence detection was conducted in the separation channel [80]. Fluorescence detection may be adequate for most applications. However, mass spectrometry (MS) has become increasingly important as an analysis tool particularly in the proteomics area because of its sensitivity. Over the last decade, the interfaces between microfluidic chips for analyte separation, electrospray ionization (ESI) ion sources, and mass spectrometers (MS) have evolved towards the successful integration of these components. The advantage of integration is the ability to achieve leak-free seals and zero dead volumes, especially between the separation channel and the ESI needle, which are critical in the separation sciences. Costs are also minimal if the chips are batch-fabricated [81]. A prime example of such a microsystem has been developed by Y. C. Tai and coworkers. The chips, shown in Fig. 11.5, were fabricated on a silicon wafer using standard photolithography and parylene as the main structural material. Three electrolysis-based electrochemical pumps were employed for loading the sample and delivering the solvent gradient. Platinum electrodes deliver current to the pumps and establish the electrospray potential. The ESI nozzle is coupled to a mass spectrometer (not shown) such that liquid chromatographytandem mass spectrometry (LC-MS/MS) could be performed [82]. The fabricated structures were able to withstand pressures in excess of 250 psi. Gradient elution through the column was performed at a flow rate of 80 nL/min. Compared to the analysis of the same sample using a commercial nanoflow LC system, the chromatographic resolution was comparable, and the total cycle time was significantly reduced because of the minimal volume between the pumps and the column. This device illustrates the potential of mass-produced, low-cost microfluidic systems capable of performing LC separations.
398
Lee and Lin
Fig. 11.5 Photograph and side view diagram of integrated LC chip containing a gradient pump, injector, mixer, reversed-phase column, electrodes, and ESI nozzle. Materials used include parylene, silicon, and PDMS. Once the chip is complete, a 5mm-thick Ultem cover is sealed on top. Chambers are machined into the cover (Reprinted with permission from [82]. Copyright 2005 American Chemical Society).
Recently Eksigent Technologies has taken it one step further by incorporating microfluidic flow control to reproducibly introduce the sample into the separation column without any losses [83]. Most chip-ESI-MS studies deal with the same problem, namely, how to introduce a nL sized sample reproducibly into the separation channel/column. Eksigent’s NanoLC system was developed for nanoscale high performance liquid chromatography (HPLC) for proteomics and high throughput HPLC for drug discovery. The NanoLC systems employ Eksigent's Microfluidic Flow Control (MFC) technology, which creates rapid, reproducible, low-flow-rate binary gradients using continuous flow measurement in combination with a con-
Current and Future Trends
399
trolled pressure source. Measurement of temperature allows the system to adjust for viscosity changes over extended run times and produces excellent retention time reproducibility day-to-day and month-to-month, for improved identification of low-abundance proteins. Users can adjust flow rates in real time to deliver accurate, consistent and precise analyses, runafter-run. In addition to liquid-phase analytes, microfluidics has played an important role in the analysis of gas-phase analytes. Microscale gas chromatographs (GCs) have followed a similar development path as liquid chromatographs towards fully integrated microsystems. A “first generation” integrated system was demonstrated by Lu and coworkers at the University of Michigan [83]. The determination of an 11-vapor mixture of typical indoor air contaminants in less than 90s was demonstrated using a hybrid system that incorporated all of the essential analytical components of a microscale GC: a sample inlet with particulate filter, on-board calibration-vapor source, multi-stage preconcentrator, dual-column separation module, an array of microsensors for analyte recognition and quantification, and a pump and valves to direct sample flow. Separate micromachined components were connected by a microfluidic channel network fabricated by deep reactive ion etching of silicon anodically bonded to a glass cover plate. Before this work, there were no reports on a complete microsystem demonstrating the capture, separation, and detection of common environmental air contaminant mixtures in the low-ppb range. Recently Murphy and coworkers have demonstrated an extremely fast microGC, delivering separation results of a mixture of four hydrocarbons in less than 2 seconds [83]. The separation channel was fabricated using LIGA, which creates high aspect ratio channels (50um wide, 600um tall). The large sidewall surface area minimizes the effects of “pooling” in the channel corners and increases the total channel surface areas available for absorbing the gases. These two examples illustrate the great potential that microfluidics will play in the development of new generation gas-phase analysis devices. Both microscale GCs and LCs utilize the advantages of microfabrication – i.e. small size which leads to increased portability, lower operating and maintenance costs, lower sample size and power requirements, and improved separation performance by minimizing dead volume.
400
Lee and Lin
11.2.5 Microfluidics for cell biology – seeing inside the cell with molecular probes Dr. Luke Lee’s lab at the University of California at Berkeley has been developing several microfluidic platforms termed BASICs (Biological Application Specific Integrated Circuits) to study “systems biology” [84]. Their goal is to understand how cells function by both the intercellular and intracellular molecular interactions. To do this, a platform to isolate specific molecular events in singular cells as well as a platform to study an ensemble of cell-cell interactions is needed. Conventional tools to study cellular events are typically focused on a “snapshot” of the system and no dynamics information is acquired. For example, protein and DNA microarrays identify what is in the sample but reveals little if anything of the functions. Functional genomics attempts to link gene expression with cellular function yet the information is still limited to a single “state” in time without dynamic information since the temporal resolution of the methods are too low. For studying specific molecular events, the most powerful tools being developed are molecular imaging techniques inside live cells on microfluidic platforms. Utilizing microfluidics control, it is possible to position the molecules or cells of interest via either hydrodynamic focusing or electrical field focusing techniques. Specific probes to study the cellular events are being developed to decipher more dynamic information to shed light on the specific functions of the molecular events. Dr. Jeff T.-H. Wang at Johns Hopkins University has developed advanced techniques utilizing molecular beacons [85] and also a promising probe termed “Quantum DotFRET” or “QD-FRET” [86]. QD-FRET can detect single molecules due to the combination of FRET sensitivity and the molecular specificity of nucleic acid binding. Luke Lee’s group has recently developed a plasmon resonance energy transfer (PRET) probe that has tremendous potential to observe intracellular molecular events [87]. PRET occurs when the plasmon resonance peaks of the metallic nanoparticle overlap with the electronic resonance peak positions of the biomolecule. As a result, different sized Au nanoparticles correlate to different binding events making it possible to identify different molecular events inside the cell. Combined with microfluidics for sorting, positioning, and counting of cells, these nanoprobes provide the future for “experimental systems biology”.
Current and Future Trends
401
11.2.6 Microfluidics for cell biology – high throughput platforms High throughput processing of cells by microfluidics is another exciting area of research that is starting to make an impact in cell based assays, cell based therapies, and tissue engineering. High throughput processing is also required to study the ensemble of cell-cell interactions. One must be able to probe signals from the single cell as it interacts and reacts to other cells in the sample. In addition, a platform is needed to record a large number of single cells and isolate different molecular parameters as they are probed. Joel Voldman’s lab has developed a dynamic array cytometer based on DEP for single cell trapping [58]. Our lab has developed a DEP flow through trapping system that allows for the sorting of a cells based on their dielectric properties and not requiring specific biomarkers to identify the cell type [61, 88]. Luke Lee’s lab has developed a nanoliter bioreactor array platform that allows cells to be cultured in an environment that mimics physiological tissue conditions with convective and diffusive nutrient delivery mechanisms [89-91]. Dave Beebe has developed a “passive microfluidics” platform (more details can be found in Chapter 8: High throughput screening using microfluidics). Dave Beebe, Luke Lee and Klavs Jensen have written excellent reviews on the cell analysis microchip platforms [84, 92, 93]. Microfluidics can also be employed in the selection and collection of single cells from within a heterogeneous population, which is required to produce genetically engineered cell lines, develop new stem cell lines, and perform single-cell studies. Selecting and isolating a single cell from a mixed population is common practice in biomedical research. Dr. Mark Bachman, Dr. G. P. Li, and co-workers have developed a new platform for the selection of single live mammalian cells while the cells remain attached to their growth surface. Millions of cells are grown on arrays of microfabricated, releasable elements composed of SU-8 polymer (“cell pallets”). The individual pallets are composed of SU-8, a biocompatible negative photoresist that is easily patterned into micrometer-sized features. Single pallets within large arrays of pallets were released on demand using a single, focused, laser pulse [94, 95]. An important advantage is that individual elements of the array are indexed so that each cell has a unique address and can be followed over time prior to its selection (see Fig. 11.6). Combined with microfluidics, the pallets can be routed to various chambers for culturing and analysis on the same substrate at the array.
402
Lee and Lin
Fig. 11.6 Schematic of cell sorting using micropallet array. A single cell attached to Su-8 substrate is released from the array. This cell is then cultured separately. With microfluidics, the pallet can be routed to a separate culture chamber in a microfluidic circuit. (Reproduced from [95] with kind permission of Springer Science+Business Media.)
The use of microfluidic devices for performing high throughput studies in cell biology holds considerable promise when compared to traditional cell culture systems (Petri dishes, mutliwell plates, etc.). Dr. David Beebe has been leading efforts to incorporate microfluidics into standard microtiter plate infrastructure by attempting to overcome compatibility barriers that hinder the application of microfluidic devices in life sciences. His group has developed microchannel cell culture platforms having both operational (compatible with plate readers and robotic pipetting) and performance (lower detection limits, controlled microenvironment) advantages. To date they have demonstrated parallel rapid growth assays and immunocytochemistry on epithelial cells in the microchannel arrays [96]. These platforms have also proven superior in terms of sample washing and experimental error in treatment concentrations. Mathematical results suggest microchannels can provide 10 or more times the treatment precision of standard wells for volume ratios typical of high-throughput screening. [97]. These microchannel platforms have also proven useful in performing parallel studies on the effects of secreted endogenous factors critical to cellto-cell signaling in processes such as wound healing and angiogenesis.
Current and Future Trends
403
Traditional multiwell plates have limited spatial and temporal control. For example, multiwell plates have a large air–liquid interface that gives rise to convective flows that continually mix the fluid disrupting the local diffusion-based accumulation. Simple microchannels provide a more controlled microenvironment that can be used to study secreted factor effects. It was found that when there are sufficient endogenous growth factors (e.g. high cell density, or fast production), the microchannels allow these factors to accumulate within entire channels more rapidly than in the wells due to the scale. When endogenous factors are scarce (e.g. low seeding density, or slow production), the microchannels allow the retention of signals locally [98]. Dr. Hugh Fan at the University of Florida has developed a novel method for toxin detection using protein expression as the assay. One of the mechanisms by which toxins cause toxic effects is to inhibit protein synthesis in cells. Thus, the Fan group developed an assay that monitors protein synthesis in an array of microwells [99]. The wells hold 13uL of fluid and are milled from acrylic. Proof of concept was demonstrated in vitro using three proteins: green fluorescent protein (GFP), chloramphenicol acetyltransferase (CAT), and luciferase. Differential inhibitory effects of two toxin simulants, tetracycline (TC) and cycloheximide (CH), on the protein expression yield of the three chosen proteins were observed, providing a unique response pattern of the array device for each toxin stimulant. Also, calibration curves were derived from the relationship between the protein yield and the amount of a toxin present. These results suggest that it is feasible to detect and identify known toxins based on the mechanisms of toxin actions. In principle, an unknown agent can be identified by comparing the response pattern with signatures of known agents in an established database.
11.3 The Future – Seamless and Ubiquitous MicroTAS With the many innovations and the ever growing number of researchers in the field, microfluidics is revolutionizing the way we handle “fluids” in almost any field one can imagine including biotechnology, biomedicine, chemical synthesis, environmental monitoring, military, covert operations/intelligence, cosmetics, advanced materials, alternate energy sources, green technologies, and toys, etc. The following are what we believe to be the high impact, new, and enabling capabilities that microfluidics will deliver.
404
Lee and Lin
Autonomous integrated fluidic chips (ai-flips) – Already, several groups have demonstrated the integration and miniaturization of many key functions for genetic analysis [100, 101]. Many platforms, including ones mentioned earlier in the chapter, provide the basis of total integration and realization of a micro TAS. For the future this holds the promise of truly portable and compact systems that can bring the biochemistry lab to the user. It will enable the delivery of healthcare to the masses and greatly facilitate the “global health” efforts being spearheaded by the Bill and Melinda Gates Foundation [102]. It will allow stand alone, autonomous, and distributed monitors for public health and surveillance for biothreats. The aging population will have many health kits to provide preventive medicine (personalized medicine or assistive medicine). Medical checkups can be performed daily. Healthcare costs will be reduced due to the availability of microfluidics instruments. Chronic disease patients can have home diagnostics systems and high risk patients (e.g. post surgical) can also enjoy the comfort of being at home knowing that lab-on-a-chip devices will be monitoring and pre-warning patients with high risks of infection, blood clots. Personalized medicine with genetic information to assist in the prescription and administration of drugs and treatments will greatly increase the efficacy and safety of the future of medicine enabled by microfluidic home test machines. With the recent environmental and national security threats, microfluidics can assist in the ubiquitous monitoring of infectious disease agents, biological engineered pathogens, and toxic substances in the environment. The ai-flips will be distributed in public areas such as buildings, airplanes, subways, parks, stadiums, schools, and malls, etc. to collect biochemical information and detect the onset of major infectious diseases or the early signs of a biological and chemical terrorism attack. For microfluidics to truly make an impact in the commercial sector with products that serve key applications, it is imperative that Standard Integration Modular Platforms (SIMPs) are developed allowing for different platforms and devices to connect as modules with standard size, power, and performance requirements. This makes it important that the microfluidic communities, including academia, industry, and government come together and agree upon these SIMPs and enforce them so that commercialization hurdles are eliminated one by one. The Micro/nano Fluidic Fundamentals Focus (MF3) Center [103] headquartered at UC Irvine and cosponsored by DARPA and industrial members is one example of an alli-
Current and Future Trends
405
ance that will play a critical role in the development of SIMPs to speed up the transition from academic research to commercialization. There are other biotechnology paradigms including artificial cells, synthetic “hybrid” biomolecules, systems biology, and synthetic biomaterials that will require the development of sophisticated microfluidic platforms down to the true nanometer scale dimensions. Future products that reflect the “inner space” ship [104] will be in the form of artificial organs (e.g. artificial kidney or pancreas [105, 106]) and sophisticated nanoparticles that track and target diseases much like what was envisioned decades ago in the movie “Fantastic Voyage”[107].
References 1. Manz A, Graber N, Widmer HM (1990) MINIATURIZED TOTAL CHEMICAL-ANALYSIS SYSTEMS - A NOVEL CONCEPT FOR CHEMICAL SENSING. Sensors and Actuators B-Chemical 1:244-248. 2. Hanson K, Nicolau Jr. DV, Filipponi L, Wang L, Lee AP, Nicolau DV (2006) Fungi Use Efficient Algorithms for the Exploration of Microfluidic Networks. Small 2:1212-1220. 3. Xia Y, Whitesides GM (1998) Soft Lithography. Angewandte Chemie International Edition 37:550-575. 4. Kupka RK, Bouamrane F, Cremers C, Megtert S (2000) Microfabrication: LIGA-X and applications. Applied Surface Science 164:97-110. 5. Fedosejevs R, Argument M, Sardarli A, Kirkwood SE, Holenstein R, Tsui YY (2003) Laser micromachining for microfluidic, microelectronic and MEMS applications. In: MEMS, NANO and Smart Systems, 2003. Proceedings. International Conference on. 6. Prashant M, Sarkar A, Lal R (2005) Facile fabrication of microfluidic systems using electron beam lithography. Lab on a Chip 6:310-315. 7. Xiaogang L, Gip S, Mirkin CA (2003) Surface and Site-Specific Ring-Opening Metathesis Polymerization Initiated by Dip-Pen Nanolithography. Angewandte Chemie 115:4933-4937. 8. Becker H, Gartner C (2000) Polymer microfabrication methods for microfluidic analytical applications. Electrophoresis 21:12-26. 9. Harrison DJ, Fluri K, Seiler K, Fan Z, Effenhauser CS, Manz A (1993) Micromachining a Miniaturized Capillary Electrophoresis-Based Chemical Analysis System on a Chip. Science 261:895-897. 10. http://www.fluidigm.com/. 11. van Dam RM (2005) Solvent-resistant elastomeric microfluidic devices and applications. Ph.D Dissertation.California Institute of Technology.
406
Lee and Lin
12. Siegel AC, Shevkoplyas SS, Weibel DB, Bruzewicz DA, Martinez AW, Whitesides GM (2006) Cofabrication of Electromagnets and Microfluidic Systems in Poly(dimethylsiloxane). Angew Chem Int Ed 45:6877-6882. 13. Siegel AC, Bruzewicz DA, Weibel DB, Whitesides GM (2007) Microsolidics: Fabrication of three-dimensional metallic microstructures in poly(dimethylsiloxane). Advanced Materials 19:727-733. 14. Lemoff AV, Lee AP (2000) An AC magnetohydrodynamic micropump. Sensors and Actuators B-Chemical 63:178-185. 15. Shaikh KA, Ryu KS, Goluch ED, Nam JM, Liu J, Thaxton CS, Chiesl TN, Barron AE, Lu Y, Mirkin CA, Liu C (2005) A modular microfluidic architecture for integrated biochemical analysis. In: Proceedings of the National Academy of Sciences 102:9745-9750. 16. Teh SY, Lin R, Hung LH, Lee AP (2008) Droplet Microfluidics. Lab on a Chip 8:198-220. 17. Jensen K, Lee AP (2004) The Science & Applications of Droplets in Microfluidic Devices. Lab on a Chip 4:31N-32N. 18. Belder D (2005) Microfluidics with droplets. Angewandte ChemieInternational Edition 44:3521-3522. 19. Gunther A, Jensen KF (2006) Multiphase microfluidics: from flow characteristics to chemical and materials synthesis. Lab on a Chip 6:1487-1503. 20. Song H, Chen DL, Ismagilov RF (2006) Reactions in droplets in microflulidic channels. Angewandte Chemie-International Edition 45:7336-7356. 21. Kobayashi I, Uemura K, Nakajima M (2007) Formulation of monodisperse emulsions using submicron-channel arrays. Colloids and Surfaces aPhysicochemical and Engineering Aspects 296:285-289. 22. Fair RB (2007) Digital microfluidics: is a true lab-on-a-chip possible? Microfluidics and Nanofluidics 3:245-281. 23. Nisisako T, Torii T (2007) Formation of biphasic Janus droplets in a microfabricated channel for the synthesis of shape-controlled polymer microparticles. Advanced Materials 19:1489-1493. 24. Nisisako T, Okushima S, Torii T (2005) Controlled formulation of monodisperse double emulsions in a multiple-phase microfluidic system. Soft Matter 1:23-27. 25. Utada AS, Lorenceau E, Link DR, Kaplan PD, Stone HA, Weitz DA (2005) Monodisperse double emulsions generated from a microcapillary device. Science 308: 537-541. 26. Xu JH, Li SW, Wang YJ, Luo GS (2006) Controllable gas-liquid phase flow patterns and monodisperse microbubbles in a microfluidic T-junction device. Applied Physics Letters 88:133506. 27. Oh HJ, Kim SH, Baek JY, Seong GH, Lee SH (2006) Hydrodynamic microencapsulation of aqueous fluids and cells via 'on the fly' photopolymerization. Journal of Micromechanics and Microengineering 16:285-291. 28. Tan YC, Hettiarachchi K, Siu M, Pan YP, Lee AP (2006) Controlled microfluidic encapsulation of cells, proteins, and microbeads in lipid vesicles. Journal of the American Chemical Society 128:5656-5658.
Current and Future Trends
407
29. Choi CH, Jung JH, Rhee Y, Kim DP, Shim SE, Lee CS (2007) Generation of monodisperse alginate microbeads and in situ encapsulation of cell in microfluidic device. Biomedical Microdevices 9:855-862. 30. Lau BTC, Baitz CA, Dong XP, Hansen CL (2007) A complete microfluidic screening platform for rational protein crystallization. Journal of the American Chemical Society 129:454-455. 31. Zheng B, Roach LS, Ismagilov RF (2003) Screening of protein crystallization conditions on a microfluidic chip using nanoliter-size droplets. Journal of the American Chemical Society 125:11170-11171. 32. Zheng B, Gerdts CJ, Ismagilov RF (2005) Using nanoliter plugs in microfluidics to facilitate and understand protein crystallization. Current Opinion in Structural Biology 15:548-555. 33. Gerdts CJ, Tereshko V, Yadav MK, Dementieva I, Collart F, Joachimiak A, Stevens RC, Kuhn P, Kossiakoff A, Ismagilov RF (2006) Time-controlled microfluidic seeding in nL-volume droplets to separate nucleation and growth stages of protein crystallization. Angewandte Chemie, International Edition 45:8156-8160. 34. Hsieh ATH, Pan JH, Lin YA, Lee AP (2006) Evaluation of Hybridization Kinetics of Molecular Beacons Using Picoliter Microfluidic Droplet with Millisecond Resolution. In: The Tenth International Conference on Miniaturized Systems for Chemistry and Life Sciences Tokyo, Japan:1321-1322. 35. Wang W, Li ZX, Luo R, Lu SH, Xu AD, Yang YJ (2005) Droplet-based micro oscillating-flow PCR chip. Journal of Micromechanics and Microengineering 15:1369-1377. 36. Beer NR, Hindson BJ, Wheeler EK, Hall SB, Rose KA, Kennedy IM, Colston BW (2007) On-Chip, Real-Time, Single-Copy Polymerase Chain Reaction in Picoliter Droplets. Anal Chem 79:8471-8475. 37. Abraham S, Jeong EH, Arakawa T, Shoji S, Kim KC, Kim I, Go JS (2006) Microfluidics assisted synthesis of well-defined spherical polymeric microcapsules and their utilization as potential encapsulants. Lab on a Chip 6:752756. 38. Chan EM, Alivisatos AP, Mathies RA (2005) High-temperature microfluidic synthesis of CdSe nanocrystals in nanoliter droplets. JOURNAL OF THE AMERICAN CHEMICAL SOCIETY 127:13854-13861. 39. Hung LH, Choi KM, Tseng WY, Tan YC, Shea KJ, Lee AP (2006) Alternating droplet generation and controlled dynamic droplet fusion in microfluidic device for CdS nanoparticle synthesis. LAB ON A CHIP 6:174-178. 40. Hung LH, Lee AP (2007) Microfluidic Devices for the Synthesis of Nanoparticles and Biomaterials. Journal of Medical and Biological Engineering, 27:16. 41. Liu K, Ding HJ, Liu J, Chen Y, Zhao XZ (2006) Shape-controlled production of biodegradable calcium alginate gel microparticles using a novel microfluidic device. LANGMUIR 22:9453-9457. 42. Flanagan LA, Lu J, Wang L, Marchenko SA, Jeon NL, Lee AP, Monuki ES (2007) Unique Dielectric Properties Distinguish Stem Cells and Their Differentiated Progeny. Stem Cells 26:656-665.
408
Lee and Lin
43. Sakai S, Hashimoto I, Kawakami K (2006) Usefulness of flow focusing technology for producing subsieve-size cell enclosing capsules: Application for agarose capsules production. Biochemical Engineering Journal 30:218-221. 44. Sakai S, Hashimoto I, Kawakami K (2007) Agarose-gelatin conjugate for adherent cell-enclosing capsules. Biotechnology Letters 29:731-735. 45. Hettiarachchi K, Talu E, Longo ML, Dayton PA, Lee AP (2007) On-chip generation of microbubbles as a practical technology for manufacturing contrast agents for ultrasonic imaging. LAB ON A CHIP 7:463-468. 46. Dayton PA, Ferrara KW (2002) Targeted imaging using ultrasound. Journal of Magnetic Resonance Imaging 16:362-377. 47. Lanza GM, Wickline SA (2001) Targeted ultrasonic contrast agents for molecular imaging and therapy. Progress in Cardiovascular Diseases 44:13-31. 48. Wei K, Kaul S (1997) Recent advances in myocardial contrast echocardiography. Current Opinion in Cardiology 12:539-546. 49. Forsberg F, Liu JB, Merton DA, Rawool NM, Goldberg BB (1995) PARENCHYMAL ENHANCEMENT AND TUMOR VISUALIZATION USING A NEW SONOGRAPHIC CONTRAST AGENT. Journal of Ultrasound in Medicine 14:949-957. 50. Goldberg BB, Raichlen JS, Forsberg F (2001) Ultrasound Contrast Agents. London: Dunitz Ltd. 51. Lee AP, Collins J, Lemoff AV (2006) A Multi-Functional Micro Total Analysis System (uTAS) Platform for Transport and Sensing of Biological Fluids using Microchannel Parallel Electrodes In: Biomolecular Sensing, Processing and Analysis, R. Bashir and S. Wereley, Editors.Springer: New York City:135-158. 52. Wang L (2007) Microfluidic Systems based on Sidewall Microelectrodes for Lateral Cell Manipulation and Analysis, Ph.D Dissertation. University of California, Irvine: Irvine. p. 149. 53.Markx GH, Rousselet J, Pethig R (1997) DEP-FFF: field-flow fractionation using non-uniform electric fields. J Liquid Chromatogr Relat Technol 20:28572872. 54. Huang Y, Wang XB, Becker FF, Gascoyne PR (1997) Introducing dielectrophoresis as a new force field for field-flow fractionation. Biophys J 73:11181129. 55. Fiedler S, Shirley SG, Schnelle T, Fuhr G (1998) Dielectrophoretic sorting of particles and cells in a microsystem. Anal Chem 70:1909-1915. 56. Iliescu C, Xu GL, Samper V, Tay FEH (2005) Fabrication of a dielectrophoretic chip with 3D silicon electrodes. Journal of Micromechanics and Microengineering 15:494-500. 57. Voldman J, Toner M, Schmidt MA (2003) Design and analysis of extruded quadrupolar dielectrophoretic traps. J Electrostat 57:69-90. 58. Voldman J, Toner M, Gray ML, Schmidt MA (2002) A MicrofabricationBased Dynamic Array Cytometer. Analytical Chemistry 74:3984-3990. 59. Park BY, Madou MJ (2005) 3-D electrode designs for flow-through dielectrophoretic systems. ELECTROPHORESIS 26:3745-3757.
Current and Future Trends
409
60. Holmes D, Green NG, Morgan H (2003) Microdevices for dielectrophoretic flow - through cell separation. Engineering in Medicine and Biology Magazine, IEEE 22:85-90. 61. Wang L, Flanagan L, Lee AP (2007) Side-Wall Vertical Electrodes for Lateral Field Microfluidic Applications. J Microelectromechanical Systems 16:454461. 62. Wang L, Flanagan LA, Jeon NL, Monuki E, Lee AP (2007) Dielectrophoresis switching with vertical sidewall electrodes for microfluidic flow cytometry. Lab Chip7:1114-1120. 63. Bansal T, Chang MP, Maharbiz MM (2007) A class of low voltage, elastomermetal 'wet' actuators for use in highdensity microfluidics. Lab on a Chip 7:164-166. 64. Xie J, Shih J, Lin QA, Yang BZ, Tai YC (2004) Surface micromachined electrostatically actuated micro peristaltic pump. Lab on a Chip 4:495-501. 65. Unger MA, Chou HP, Thorsen T, Scherer A, Quake SR (2000) Monolithic microfabricated valves and pumps by multilayer soft lithography. Science 288:113-116. 66. Vezenov DV, Mayers BT, Conroy RS, Whitesides GM, Snee PT, Chan Y, Nocera DG, Bawendi MG (2005) A low-threshold, high-efficiency microfluidic waveguide laser. JOURNAL OF THE AMERICAN CHEMICAL SOCIETY 127:8952-8953. 67. Tang SKY, Mayers BT, Vezenov DV, Whitesides GM (2006) Optical waveguiding using thermal gradients across homogeneous liquids in microfluidic channels. Applied Physics Letters 88:061112. 68. Wolfe DB, Conroy RS, Garstecki P, Mayers BT, Fischbach MA, Paul KE, Prentiss M, Whitesides GM (2004) Dynamic control of liquid-core/liquidcladding optical waveguides. In: Proceedings of the National Academy of Sciences of the United States of America 101:12434-12438. 69. Martinez AW, Phillips ST, Butte MJ, Whitesides GM (2007) Patterned paper as a platform for inexpensive, low-volume, portable bioassays. Angewandte Chemie-International Edition 46:1318-1320. 70. Fodor SPA, Read JL, Pirrung MC, Stryer L, Lu AT, Solas D (1991) LIGHTDIRECTED, SPATIALLY ADDRESSABLE PARALLEL CHEMICAL SYNTHESIS. Science 251:767-773. 71. Liu R, Lee AP eds. Integrated Biochips for DNA Analysis.(2007) Biotechnology Intelligence Unit. Landes Bioscience: Austin, Texas:199. 72. Liu RH, Lodes MJ, Nguyen R, Siuda T, Slota M, Fuji HS, McShea A (2006) Validation of a fully integrated microfluidic array device for influenza A subtype identification and sequencing. Analytical Chemistry 78:4184-4193. 73. Liu RH, Nguyen T, Schwarzkopf K, Fuji HS, Petrova A, Siuda T, Peyvan K, Bizak M, Danley D, McShea A (2006) Fully integrated miniature device for automated gene expression DNA microarray processing. Analytical Chemistry 78:1980-1986. 74. Chang-Yen DA, Myszka DG, Gale BK (2006) A novel PDMS microfluidic spotter for fabrication of protein chips and microarrays. Journal of Microelectromechanical Systems 15:1145-1151.
410
Lee and Lin
75. Zhang ZX, Shen Z, Zhao H, Li B, Song SP, Hu J, Lin BC, Li MQ (2005), Microarrays of DNA and protein integrated on microfluidic chip. Acta Chimica Sinica 63:1743-1746. 76. Davies DH, Liang XW, Hernandez JE, Randall A, Hirst S, Mu YX, Romero KM, Nguyen TT, Kalantari-Dehaghi M, Crotty S, Baldi P, Villarreal LP, Felgner PL (2005) Profiling the humoral immune response to infection by using proteome microarrays: High-throughput vaccine and diagnostic antigen discovery. In: Proceedings of the National Academy of Sciences of the United States of America 102:547-552. 77. Nicewarner-Pena SR, Freeman RG, Reiss BD, He L, Pena DJ, Walton ID, Cromer R, Keating CD, Natan MJ (2001) Submicrometer metallic barcodes. Science 294:137-141. 78. Brunker SE, Cederquist KB, Keating CD (2007) Metallic barcodes for multiplexed bioassays. Nanomedicine 2:695-710. 79. Pregibon DC, Toner M, Doyle PS (2007) Multifunctional encoded particles for high-throughput biomolecule analysis. Science 315:1393-1396. 80. Fuentes HV, Woolley AT (2007) Electrically actuated, pressure-driven liquid chromatography separations in microfabricated devices. Lab on a Chip 7:1524-1531. 81. Koster S, Verpoorte E (2007) A decade of microfluidic analysis coupled with electrospray mass spectrometry: An overview. Lab on a Chip 7:1394-1412. 82. Xie J, Miao YN, Shih J, Tai YC, Lee TD (2005) Microfluidic platform for liquid chromatography-tandem mass spectrometry analyses of complex peptide mixtures. Analytical Chemistry 77:6947-6953. 83. Eksigent web page. http://www.eksigent.com. 84. Breslauer DN, Lee PJ, Lee LP (2006) Microfluidics-based systems biology. Molecular Biosystems 2:97-112. 85. Ho YP, Kung MC, Yang S, Wang TH (2005) Multiplexed hybridization detection with multicolor colocalization of quantum dot nanoprobes. Nano Letters 5:1693-1697. 86. Ho YP, Chen HH, Leong KW, Wang TH (2006) Evaluating the intracellular stability and unpacking of DNA nanocomplexes by quantum dots-FRET. Journal of Controlled Release 116:83-89. 87. Liu GL, Long YT, Choi Y, Kang T, Lee LP (2007) Quantized plasmon quenching dips nanospectroscopy via plasmon resonance energy transfer. Nature Methods 4:1015-1017. 88. Wang L, Marchenko S, Jeon NL, Monuki ES, Flanagan LA, Lee AP (2006) Lateral Cell Separation in Microfluidic Channel by 3D Electrode Dielectrophoresis. In: The Tenth International Conference on Miniaturized Systems for Chemistry and Life Sciences (uTAS 2006). Transducers Research Foundation. 89. Lee PJ, Hung PJ, Rao VM, Lee LP (2006) Nanoliter scale microbioreactor array for quantitative cell biology. Biotechnology and Bioengineering 94:5-14. 90. Di Carlo D, Lee LP (2006) Dynamic single-cell analysis for quantitative biology. Analytical Chemistry 78:7918-7925. 91. Di Carlo D, Wu LY, Lee LP (2006) Dynamic single cell culture array. Lab on a Chip 6:1445-1449.
Current and Future Trends
411
92. El-Ali J, Sorger PK, Jensen KF (2006) Cells on chips. Nature 442:403-411. 93. Yu HM, Meyvantsson I, Shkel IA, Beebe DJ (2005) Diffusion dependent cell behavior in microenvironments. Lab on a Chip 5:1089-1095. 94. Salazar GT, Wang YL, Young G, Bachman M, Sims CE, Li GP, Allbritton NL (2007) Micropallet arrays for the separation of single, adherent cells. Analytical Chemistry 79:682-687. 95. Sims CE, Bachman M, Li GP, Allbritton NL (2007) Choosing one from the many: Selection and sorting strategies for single adherent cells. Analytical and Bioanalytical Chemistry 387:5-8. 96. Yu HM, Alexander CM, Beebe DJ (2007) A plate reader-compatible microchannel array for cell biology assays. Lab on a Chip 7:388-391. 97. Warrick J, Meyvantsson I, Ju JI, Beebe DJ (2007) High-throughput microfluidics: improved sample treatment and washing over standard wells. Lab on a Chip 7:316-321. 98. Yu HM, Alexander CM, Beebe DJ (2007) Understanding microchannel culture: parameters involved in soluble factor signaling. Lab on a Chip 7:726730. 99. Mei Q, Fredrickson CK, Jin SG, Fan ZH (2005) Toxin detection by a miniaturized in vitro protein expression array. Analytical Chemistry 77:5494-5500. 100. Liu RH, Yang JN, Lenigk R, Bonanno J, Grodzinski P (2004) Self-contained, fully integrated biochip for sample preparation, polymerase chain reaction amplification, and DNA microarray detection. Analytical Chemistry 76:18241831. 101. Burns MA, Mastrangelo CH, Sammarco TS, Man FP, Webster JR, Johnson BN, Foerster B, Jones D, Fields Y, Kaiser AR, Burke DT (1996) Microfabricated structures for integrated DNA analysis.In: Proceedings of the National Academy of Sciences 93:5556-5561. 102. Bill and Melinda Gates Foundation. Available from: http://www.gatesfoundation.org/. 103. Lee A, Lin G Micro/nano Fluidics Fundamentals Focus (MF3) Center: Irvine, CA. 104. Dante J, (1987) Inner Space Amblin Entertainment, Inc.: USA. 105. Yang TH, Lee CJ, Chu IM, Hwang JC, Yang TL (1999) The Removal of Beta-2-microglobulin by an Immunoadsorption Wall. Artificial Organs 23:131-138. 106. Wise KD (2007) Integrated sensors, MEMS, and microsystems: Reflections on a fantastic voyage. Sensors and Actuators A: Physical 136:39-50. 107. Fleischer R, Fantastic Voyage. 1966, Twentieth Century-Fox Film Corporation: USA.
Index
A Abrasive powder machining, 51, 55 Acoustophoresis, 199 Adhesive bonding, 58 Anodic bonding, 57 B Biocompatible material, 97 Biodefense, 325, 326, 404 Biodegradable materials, 78, 228, 390 Biological Autonomous Networked Detectors (BAND), 327, 353 BioWatch, 327 Bond number, 14, 15 Braille display systems, 205, 206 Bulk micromachining, 39 C Cantilevers, 18, 292, 293
Capillary electrophoresis, 122, 131, 134, 168, 169, 255, 283, 284, 354, 357, 358, 395 Carbon nanotubes, 288 Casting, 63 Cell culture, 187, 200, 201, 202, 204, 245, 250, 275, 401, 402 Cell interactions, 248, 249 Cell ligand adhesion, 195 Cell separation, 186, 191, 193, 194, 195, 198, 199, 401 Cell sorting, 24, 192, 252, 253, 396, 401 Cellular analysis, 247, 258 Cellular assays, 188, 208, 209, 396, 400 Chemotaxis assay, 211 Computational fluid dynamic modeling, 236 Concentration methods, 175, 176, 177, 195, 280, 281, 286, 287, 296, 357 Continuum approximation, 10
414 D Dead volume, 97, 98, 99 Deep Reactive Ion Etching (DRIE), 44 Dielectrophoresis (DEP), 253, 285, 286, 296, 337, 389, 393, 401 Diffusion, 12, 13, 15, 73 Digital microfluidics, 257, 390 DNA sequencing, 136, 138, 139, 140, 144, 148, 356, 357 Droplet microfluidics, 14, 15, 257, 389 Drug delivery, 24, 109, 390 Drug screening, 202, 258, 259 Dynamic coating, 126, 134, 135, 136, 169 E Electrochemical discharge machining (ECDM), 56 Electrochemilumnescence, 344 Electroosmotic flow, 5, 16, 19, 128, 134, 284 Electrophoresis, 3, 117, 119, 122, 123, 125, 137, 139, 145, 166, 168, 169, 170 Electrophoresis arrays, 137, 138, 255, 257, 354 Electrospray ion source, 103, 104, 105, 106, 107, 108, 109, 180, 345, 397, 398 Electrowetting, 15, 257 Entropic trap, 132, 133 Enzyme-linked immunosorbent assay (ELISA), 298, 299, 342 Exponential amplification reaction, 351, 352 F Filtration, 280, 281 Flow cytometry, 291, 292 Fluidic interconnects, 39, 99, 364 Fluorescence activated cell sorting (FACS), 192, 193, 194, 252
Fusion bonding, 57 G Gas chromatography, 3, 399 Glass Based Materials, 49 H H-filter, 11, 13, 14 High throughput cell culture, 245, 246, 402 High throughput screening, 209, 242, 245, 246, 255, 259, 390, 401 Hot embossing, 19, 61 Hydrogel, 72, 232, 390 I Immunoassays, 175, 290, 297, 298, 299, 333, 339, 341, 342, 343, 344 Immunomagnetic separations, 194, 282, 339, 340, 341 Injection molding, 19, 21, 60 Integrated system, 7, 8, 142, 178, 306, 358, 361, 363, 399, 404 Integration electrodes, 392 Electrodes, 287, 393 Optics, 393 Isoelectric focusing, 17, 171 Isolated channel networks, 249, 251 Isotachophoresis (ITP), 17, 168, 357 J Joint Biological Agent Identifiation and Diagnostic System (JBAIDS), 328, 329 Joint Biological Point Detection System (JBPDS), 329 Joule heating, 6, 18 L Laminar flow, 5, 6, 10, 11, 15, 23, 196 Lateral flow assays, 333, 341, 342
415 Leakage tests, 97, 98 Liquid chromatography, 3, 4, 172, 173, 175, 336, 345, 397, 398 Liver-on-a-chip, 202 Loop-mediated isothermal amplification, 351, 352 M Magnetic cell separation, 281 Magnetic force assay, 344 Metabolite detection, 294, 296 Micellar electrokinetic chromatography (MEKC), 173 Microarrays Cells, 202, 205, 208 DNA, 118, 119, 149, 350, 352, 353, 395, 396 Protein, 346, 395 Microbubble generation, 390 Microfabrication-Glass, 51 Microfabrication-Polymer, 59 Microfabrication-Silicon, 39 Microneedles, 47, 48, 109 Modular system integration, 8 Molecular imaging, 400 Monoliths, 74, 75, 174 Multidimensional protein separation, 178 N Nanogel, 128, 138 Nanoparticle generation, 390 Nucleic acid purification, 337, 338 O Ogston model, 125 P Paper microfluidics, 394, 395 Paraffin, 76 Paramagnetic beads, 194, 334, 340 Parylene, 64 Pathogen detection, 143, 274, 277, 289, 290, 293, 294, 301
Pathogen isolation, 278, 281, 283, 284, 286, 287 Photo-definition of glass, 54 Photolithography, 52, 53, 231 Point of care (POC), 7, 276 Poiseuille flow, 11 Polyacrylamide cross-linked, 127, 170 Linear, 127, 170 Polycarbonate (PC), 60, 124 Polydimethylsiloxane (PDMS), 69, 226, 279, 304, 388 Polyethylene oxide (PEO), 129 Polyimide, 71 Polymerase chain reaction (PCR), 8, 140, 254, 297, 301, 302, 332, 346, 348, 349, 350 Polymethylmethacrylate (PMMA), 60 Poly-N, N-dimethylacrylamide (PDMA), 128 Polysaccharide, 127 Polyvinyl alcohol (PVA), 129 Polyvinyl pyrrolidone (PVP), 129 Proteomics, 167, 345 Pulmonary airways, 200 Pumps, 16 Q Quantitative polymerase chain reaction (qPCR), 332, 349, 350 R Reactive ion etching (RIE), 42 Reptation theory, 125 Reynolds number, 5, 6, 10 Rolling circle amplification, 351, 352 S Sample preparation, 139, 175, 335, 336 Sanger sequencing, 119 SDS-PAGE, 170, 176, 177, 178
416 Sherwood number, 12 Short tandem repeats analysis, 122 Silicon based materials, 37 Single cell analysis, 210 Single nucleotide polymorphism (SNP), 121, 356 Sperm sorting, 196 Spore disruption, 336 Stem cell, 204, 390 Stereolithography, 63 SU-8, 66, 394 Surface micromachining, 46 T Temperature control, 18, 303, 349, 350 Temperature gradient focusing (TGF), 17
Thermoresponsive polymer, 126, 129 Time resolved fluorescence assay, 344 Tissue engineering, 225, 232 Tubular chromatography, 168, 173 V Valves, 16 Vascularized tissue constructs, 225, 226, 229, 235 Viral count, 276 W Wafer bonding, 56 Waveguide, 393, 394 Wet anisotropic etching, 40
