In This Issue
Fall Issue of The Bridge on the Convergence of Engineering and the Life Sciences
October 1, 2013 Volume 43 Issue 3

Microfabrication The Interface Between Medicine and Engineering

Wednesday, September 25, 2013

Author: Stephen R. Quake

Over the past half-century the semiconductor industry developed a powerful set of manufacturing tools that enable highly parallel fabrication of electronic devices with an extraordinarily large number of integrated components. A key part of this process is the use of lithographic techniques to transfer complex patterns into semiconductor substrates, and another part is the development of approaches to manipulate the material properties of these substrates by etching, doping, and deposition. As a result, electronic technologies have enjoyed exponential growth in sophistication and complexity with corresponding decreases in cost, as captured by the famous Moore’s law. The power of this approach to manufacturing has inspired several generations of biomedical researchers, who have sought to bring the same sort of scaling to biology and medicine. In this article I discuss several examples of how this interface has been exploited.

Background

In the 1980s engineers realized that silicon was a useful material not just for its electronic properties but also for its very interesting mechanical properties. Exploitation of these properties led to the development of the field of microelectromechanical systems (MEMS) and the production of a variety of intricate miniaturized mechanical devices and systems—from gears to ratchets to mechanical oscillators, to name just a few. A number of familiar consumer devices use MEMS, such as the micromirror arrays in some video projectors and the inertial sensors in automobile air bags (as well as in mobile phones). There have also been some direct medical applications of MEMS, including in implantable devices.

The introduction of lithography—the use of photographic techniques to effect pattern transfer—was crucial in enabling the semiconductor industry to produce monolithic devices with enormous numbers of distinct components all fabricated in situ. There have been two clear and powerful extensions of this approach to the life sciences and medicine.

Molecular Lithography

Molecular lithography is used to fabricate distinct molecular species (Fodor et al. 1991). It enabled synthesis of huge arrays of both peptides and nucleic acids, and launched the biochip industry. The DNA arrays produced by this technology also helped create the genomics revolution, by creating a technology for highly parallel gene expression measurements—in effect, creating the ability to monitor global gene expression of nearly all the genes in an organism at once. The arrays are also used for high-throughput genotyping—measuring up to a million single-nucleotide polymorphisms at a time.

Molecular lithography inspired a host of other DNA array fabrication technologies, such as robotic fountain pens (Schena et al. 1995) and self-assembled bead arrays (Epstein and Walt 2003), all of which remain in use today.

“Soft” Lithography

The other major extension of semiconductor lithography has been the introduction of a suite of fabrication techniques termed “soft lithography” (Xia and Whitesides 1998). This approach led to an expansion in the variety of materials that could be microfabricated into arbitrary structures and has also enabled exquisite control over the patterning of chemical functionalities. Soft lithography inspired a wholesale move from silicon to silicone elastomers such as polydimethyl siloxane as the material of choice for biologically related device fabrication.

Another powerful impact has been in the form of microcontact printing, in which elastomeric “stamps” are used to effect pattern transfer of chemical inks onto surfaces in defined patterns. A fascinating application of this technology has been the ability to pattern the shapes that cells take when attached to surfaces, thus making it possible to explore the connection between geometry and internal cell mechanics (Chen et al. 1997).

Other Types of Machining

It is worth noting that not all methods of biomedical microfabrication are derived from the lithography paradigm of the semiconductor industry. One example of this is laser cutting–based micromachining, which is the microfabrication technology used to manufacture stents. Another example is electric discharge machining, which is used to manufacture many of the miniaturized fountain pens used in printed DNA arrays. Three-dimensional printing is rapidly emerging as a manufacturing technology with applications for microfabrication in biology and medicine, and has been reviewed recently in this journal (Lipson 2012).

The Intersection of Microfabrication and the Life Sciences

One of the most productive interfaces between microfabrication and the life sciences has been the emergence of the field of microfluidics, which is often defined as the development and application of fluid manipulation at the nanoliter scale and is now the subject of a variety of textbooks (Folch 2012; Tabeling 2010).

Microfluidics

Microfluidics had its origins in the desire of chemists to build miniaturized and portable analytical devices. Early examples include a microfabricated gas chromatograph as well as a number of microfabricated capillary electrophoresis systems. It was soon appreciated that the fluid physics governing the behavior of nanoliter volumes in microfabricated devices is quite different from one’s intuition from the macro scale. While this was initially approached as a problem to circumvent (i.e., how does one mix fluids in the absence of inertia?), it was soon embraced as a “feature,” and there are now numerous examples of devices that exploit the unique fluid physics of small-length scales to achieve performance that is not possible in conventional volumes (Squires and Quake 2005).

Applications of Microfluidics
Large-Scale Integration

Microfluidic large-scale integration (LSI) realizes the most direct analogy with semiconductor integrated circuits. This approach involves fabricating hundreds to tens of thousands of integrated micromechanical valves in a single microfluidic device, which can be connected with arbitrarily complex plumbing designs (Melin and Quake 2007). The valves can be combined into other functional units such as peristaltic pumps, multiplexed addressing schemes, rotary mixers, and reconfigurable columns.

Just as the development of LSI design rules in digital electronics freed circuit designers from having to worry about the technical details of the transistor technology they were using, so too can microfluidic LSI be used to separate design from fabrication, since it enables abstraction of functional units on several levels. This ability to abstract design rules is one of the most powerful aspects of microfluidic LSI and remains a crucial distinguishing factor from other approaches to manipulate fluids in microfluidic devices, such as electrokinetic effects and compartmentalization by two-phase emulsions, which are more akin to analogue electronics.

Protein Crystallography

One example of the power of microfluidic LSI can be found in protein crystallography, where microfluidics enables sophisticated manipulation of the nucleation and growth kinetics of protein crystal growth, which are otherwise very difficult to achieve (Hansen and Quake 2003). The fundamental physical basis for this rests on the fact that density-driven convection is negligible in nanoliter volumes—or more precisely, forces caused by convection are small relative to the viscous dissipation of the fluid. This enables implementation of a crystal growth strategy called “free interface diffusion,” to explore the solubility phase space of the protein in a manner that best mimics the crystallization process.

Microfluidic approaches to protein crystallization have the added advantage of consuming only tiny amounts of protein sample in each experiment, thus allowing hundreds of experiments with the same amount of protein required for a single conventional experiment. Since the proteins used in these experiments can be extraordinarily difficult and expensive to purify, this engineering economy of scale is an important benefit of the microfluidic approach.

Chips to screen protein crystal growth conditions outperform conventional methods by an impressive margin. These chips have been commercialized and used in structural biology labs in industry and academia around the world, and numerous structures that resisted all prior efforts were solved with the use of this chip. Many of these are related to human health or pharmaceuticals, including the Ebola virus glycoprotein, H5N1 influenza virus hemagglutinin, and an integrin binding to a fibrogen-mimetic therapeutic.

Cell and Tissue Culture

Another area in which microfluidic devices are having an enormous impact is cell and tissue culture. It is remarkable how few technological advances this field enjoyed since the invention of the petri dish. However, in recent years there has been a burst of useful cell culture innovations based on microfluidic technologies.

The novel fluid physics of small volumes can be used to create complex chemical gradients, a capacity that is useful for understanding how cells move and respond to cues from signaling molecules in their environments (Jeon et al. 2002). Microfluidics has even been used to create thermal gradients across individual fly embryos, enabling study of the molecular clocks associated with development and their dependence on temperature (Lucchetta et al. 2005). Microfluidic technology has also been used to create miniaturized cell culture devices, in formats ranging from microbial chemostats (Balagadde et al. 2005) to high-throughput mammalian cell culture (Tay et al. 2010).

Genomic Analysis

An exciting emerging area revolves around the use of microfluidic tools for single-cell genomic analysis (Kalisky et al. 2011). Microfluidic devices have been used for both gene expression analysis and for genome sequencing from single cells. In the case of gene expression analysis, it has become routine to analyze hundreds of genes per cell on hundreds to thousands of single cells per experiment. This has led to many new insights into the heterogeneity of cell populations in human tissues, especially in the areas of cancer and stem cell biology. These devices make it possible to perform “reverse tissue engineering” by dissecting complex tissues into their component cell populations, and they are also used to analyze rare cells such as circulating tumor cells.

Single-cell genome sequencing has been used to analyze the genetic properties of microbes that cannot be grown in culture—the largest component of biological diversity on the planet—as well as the recombination potential of humans by characterizing the diversity of novel genomes found in the sperm of an individual. While it is indeed possible to do single-cell genomic experiments with traditional bulk approaches, the microfluidic approach is lower cost, high throughput, and usually of higher technical quality.

Bioanalytical Chemistry

There have been many applications of microfluidics to bioanalytical chemistry; I highlight here the emergence of digital polymerase chain reaction (PCR).

It has been a long-standing challenge of measurement science to make absolute quantitation of particular nucleic acid species. The gold standard for many years has been quantitative PCR (qPCR), but for a variety of technical reasons this tends to be a relative and not absolute measurement. Digital PCR is a way of using limiting dilution to make an absolute measurement of nucleic acid quantitation, by, in effect, partitioning a sample into many subsamples, each of which contains on average less than one molecule of DNA. By applying PCR to each subsample, one can determine which of the subsamples had a molecule and which didn’t, and effectively “count molecules” by counting the number of subsamples with amplified product in them. This is tedious and impractical to do with conventional bench techniques, but is easily enabled with microfluidic automation.

Many applications of microfluidic digital PCR have been demonstrated, and government standards agencies have contemplated its use in measuring GMO contamination in the food supply.

Nanoliter Chemical Synthesis

Microfluidic large-scale integration has also been applied to develop nanoliter-scale synthetic chemistry devices. One might question the utility of performing such synthesis, since it results in very small amounts of product, but in fact there are several useful application areas. One of these is in positron emission tomography (PET), the widely used medical imaging technique.

PET requires the use of small amounts of radiopharmaceuticals, and because of the short half-lives of the positron-emitting isotopes it would be quite useful to synthesize single doses on demand. Since only tiny amounts of the radio-compound are used per dose, individual doses can be synthesized in nanoliter reactors (Lee et al. 2005).

Another area where nanoliter chemical synthesis leads to practical applications is DNA synthesis—small amounts of DNA product can always be amplified with methods such as PCR. It is possible that microfluidic gene synthesis methods will play a role in the emerging field of synthetic biology (Lee et al. 2010).

Conclusion

Microfluidics has matured well beyond its academic origins and has had significant commercial impact in a variety of realms. There are numerous small and medium-sized companies producing a wide range of microfluidic tools—one can get a sense of the scale by noting that well over a billion of the micromechanical valves that form the basis of microfluidic large-scale integration have been manufactured and shipped to customers. These companies play an important role in helping nonexpert users enjoy the benefits of microfluidic tools, and also provide a metric to gauge the scientific and economic impact of the field.

Just as the invention of the integrated circuit catalyzed the transition from vacuum tube computers to the modern computer era, it is likely that microfluidic devices will continue to provide unprecedented advances in biological automation and productivity.

References

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(5731):137–140.

Chen CS, Mrksich M, Huang S, Whitesides GM, Ingber DE. 1997. Geometric control of cell life and death. Science 276(5317):1425–1428.

Epstein J, Walt DR. 2003. Fluorescence-based fibre optic arrays: A universal platform for sensing. Chemical Society Reviews 32(4):203–214.

Fodor SPA, Read J, Pirrung M, Stryer L, Lu A, Solas D. 1991. Light-directed, spatially addressable parallel chemical synthesis. Science 251(4995):767–773.

Folch A. 2012. Introduction to BioMEMS. Boca Raton: CRC Press.

Hansen C, Quake SR. 2003. Microfluidics in structural biology: Smaller, faster…better. Current Opinion in Structural Biology 13(5):538–544.

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(8):826–830.

Kalisky T, Blainey P, Quake S. 2011. Genomic analysis at the single-cell level. Annual Review of Genetics 45:431–445.

Lee CC, Sui G, Elizarov A, Shu CJ, Shin YS, Dooley AN, Huang J, Daridon A, Wyatt P, Stout D, Kolb HC, Witte ON, Satyamurthy N, Heath JR, Phelps ME, Quake SR, Tseng HR. 2005. Multistep synthesis of a radiolabeled imaging probe using integrated microfluidics. Science 310(5755):1793–1796.

Lee CC, Snyder TM, Quake SR. 2010. A microfluidic oligonucleotide synthesizer. Nucleic Acids Research 38(8):2514–2521.

Lipson H. 2012. Frontiers in additive manufacturing: The shape of things to come. The Bridge 42(1):5–12.

Lucchetta EM, Lee JH, Fu LA, Patel NH, Ismagilov RF. 2005. Dynamics of Drosophila embryonic patterning network perturbed in space and time using microfluidics. Nature 434:1134–1138.

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.

Schena M, Shalon D, Davis RW, Brown PO. 1995. Quantitative monitoring of gene expression patterns with a complementary DNA microarray. Science 270(5235):467–470.

Squires TM, Quake SR. 2005. Microfluidics: Fluid physics at the nanoliter scale. Reviews of Modern Physics 7(3):977–1026.

Tabeling P. 2010. Introduction to Microfluidics. New York: Oxford University Press.

Tay S, Hughey JJ, Lee TK, Lipniacki T, Quake SR, Covert MW. 2010. Single-cell NF-kappaB dynamics reveal digital activation and analogue information processing. Nature 466(7303):267–271.

Xia Y, Whitesides GM. 1998. Soft lithography. Angewandte Chemie International Edition 37:550–575.

About the Author:Stephen R. Quake (NAS/NAE/IOM) is Lee Otterson Professor in the Departments of Applied Physics and Bioengineering at Stanford University, and an investigator of the Howard Hughes Medical Institute.