Profile

Professor Paul Yager

Department of Bioengineering, University of Washington, Seattle, USA

I’m married with 2 children, ages 17 and 13. Life with two teenagers is at least as challenging (and rewarding!) as all the rest combined.

I moved to Seattle in 1987, after eying it for over a decade. I first came up here before graduate school and realized that this was a place that combined all the things I wanted in a place to live—mountains AND lakes AND salt water. The ability to walk or bike to work, art, opera, theatre, great restaurants and great public schools. Oh, and a place to tie up my sailboat within a few minutes’ walk of my office on campus. Heaven!

Employment
 
Vice Chair, Department of Bioengineering, 2001–present
Co-Director, Center for Applied Microtechnology, University of Washington, 1998–2000
Professor, Department of Bioengineering, University of Washington, 1995–present
Adjunct Faculty, Chemical Engineering, University of Washington, 1993–present
Adjunct Faculty, Chemistry, University of Washington, 1990–present
Director, Microbiosensor Technology Laboratory, Washington Technology Center, 1988–1993
Associate Professor, Center for Bioengineering, University of Washington, 1987–1995
Research Chemist, Naval Research Laboratory, Washington, DC, 1982–1987
National Research Council Resident Research Associate, Naval Research Laboratory, 1980–1982
 
Educational institution/degree, years
 
Princeton University/A.B., Biochemistry, 1971–1975
University of Oregon/Ph.D., Chemistry, 1975–1980

Research

My research career has followed an arc that began as the most basic of researchers, but has led me to be very focused on applied problems. It has been a story of accumulating research tools and bits of information over the years, and finding at the appropriate point in my career that I knew how to do practical things that other people hadn’t quite yet thought of.

I was raised in New York City by a single mother who had a degree in Chemistry, and who worked as a text editor for scientific books and journals, so in a sense I never had a chance to be anything other than a scientist. Luckily enough, I actually liked science, and have always been a scientific omnivore—I read all the articles in Scientific American, and still do when I have the time. This may have been the best background for getting into microfluidics.

As a child growing up in New York City, my earliest memory of ‘real’ science was of a visit at age 8 to the laboratory of a friend of my mother who was one of the first medical researchers who had gotten his hands on an electron microscope. I will never forget the magic of being in a dark room filled with a large and spooky machine and peering at its fluorescent screen and being one of the first people to have glimpsed things that were unimaginably small. I was hooked. From the age of about 10 I was convinced that the brain was some sort of machine, and I wanted to be the person who learned how it worked. By the time I was in high school I was volunteering in the laboratory of a Prof. Yoshiaki Omura at New York Medical College who was, I much later realized, a classic Bioengineer—an MD who taught medical students AND a Ph.D. electrical engineer who also taught engineering at Manhattan College (which was, ironically, in the Bronx). By the end of high school I had been putting together electrical circuits and doing tissue culture on cardiac pacemaker cells, and knew that a life of research was for me—the sooner the better. It was partially for that reason that I turned down (to my mother’s distress) the opportunity to join the Brown 6-year medical program (that would, as she said, have meant that ‘I would always have a job’) to try the more relaxed (or so I thought) academic life of a liberal arts student at Princeton.

What followed was a process of specialization on increasingly smaller parts of the cell. My undergraduate degree at Princeton was from the short-lived but extraordinarily fertile Biochemistry Department, where I was exposed to the physical chemistry of (bio)polymers at a very high level. Since there were no laboratories working on cell membranes per se there, I found myself from my junior year working in the laboratory of Tom Spiro in Chemistry, where I learned to use Raman spectroscopy to study lipids. By the time I graduated in 1975, I was convinced that the way to understand neural membrane function was to understand the behavior of the lipids in them. From there I moved to the laboratory of Warner Peticolas in the Chemistry Department and Institute of Molecular Biology at the University of Oregon (UofO) in Eugene. This was, at the time, the only place I could have obtained a Ph.D. working with someone interested in studying biomembranes using Raman spectroscopy. As it turns out, it was another hotbed of physical biochemistry, and it was there that I developed an abiding love of proteins. My primary work, however, focused on developing Raman spectroscopy into a sensitive probe of lipid structure, particularly using selective isotopic substitution to determine structure in selected portions of the lipid molecules in different phases. Along the way there was a brief but intense diversion on bacteriorhodopsin and understanding of photocycling in that fascinating protein. After a failed attempt to use Raman to monitor the putative changes in lipid phase that may occur during the propagation of nerve impulses down nerves (that resulted in a life-long inability to eat Dungeness crab without pangs of guilt), my thesis ended up focusing on the kinetics of pressure-induced phase transitions in phospholipids. This was as far as I was to get in understanding the brain, but I had completed my journey down to the molecular level. During my Ph.D. work I was very isolated from the day-to-day issues of support of science—this was science at its purest. However, I also discovered that I loved to make instruments in the machine shop, and that I had a knack for visualizing complex 3-D shapes.

My journey to the applied side began just after graduating in 1980; that fall I joined several former UofO colleagues including Bruce Gaber at the Naval Research Laboratory in Washington, DC, which we fondly called Peticolas Lab East. My initial National Research Council postdoctoral work was aimed at developing a liposome-based blood surrogate. After my conversion to a Research Scientist at NRL, I also worked on basic research into biomembrane structure (using Raman spectroscopy) and developing liposome-based immunosensors. Another project that I started was to use stabilized lipid bilayers as the basis for receptor-based biosensors. My first patent (well, actually a ‘statutory invention registration’) was on that idea. That project eventually bifurcated, leading to the two main themes of the next 15 years of work for me.

The first major break in the work was an accident. Paul Schoen and I were looking at using polymerizable phospholipids as the basis of tough bilayers for the sensors. At this point these lipids were about 2 years old, and we had to get one particular phospholipid lipid (one with diacetylenic moieties in the hydrocarbon chains) custom-synthesized. When I subjected these unusual lipids to thermal cycling, I found that the liposomal form was unstable below the hydrocarbon chain crystallization temperature, and that the lipids converted into a hitherto unknown hollow cylindrical form of lipids we dubbed ‘tubules’. It took a few weeks to be sure this wasn’t some weird artifact, but it was a heady experience to realize that I had seen something new to the world of science. While I viewed these tubules as a biophysical oddity, I soon learned that at NRL everything had to be sold as important to the national defense, and remarkably enough, Dr. Joel Schnur, my supervisor, was able to make this case. This discovery resulted in publications, patents, and funding from several sources, including DARPA. The mentorship of Dr. Ira Skurnick of DARPA’s Defense Sciences Office was, in large part, to prove to be pivotal for the rest of my career. Meanwhile, the bilayer-based biosensor work was proceeding, and may have blossomed had I stayed there with Thomas Fare, who introduced me to the idea of using silicon microfabrication to hold the bilayer.

However productive my time at NRL, I was not constitutionally well suited to life in a government laboratory. I began a search for an academic department (any department!) that would be willing to take a risk on a Chemistry Ph.D. who was interested in odd biosensors and even odder self-assembling lipid microstructures. I didn’t fit ANY of the descriptions in the want ads in Science. Many many application letters later, I found a department just odd enough to tolerate my odd interests; when the opportunity to move to the University of Washington’s Department of Bioengineering presented itself, I jumped.

For the first time I was in an environment that had both a Medical School and a charter to do things of biomedical relevance. Even though the lipid tubule work was fun, scientifically rich, publishable and supportable, it was, in the environment of UW (the University of Washington) Bioengineering, ultimately doomed. The work continued in a few guises, including a collaborative project with Michael Gelb to develop a drug delivery system based on the self-disassembly of the microstructures. However, it eventually fell victim to my interest in sensor development and the support for that work in my department. The lipid work that did survive was related to the ideas that I had picked up from Tom Fare. I had a graduate student who was interested in silicon microfabrication, and together we created a microfabricated ‘bilayer support device’. This 16-chamber microdevice proved capable of creating and supporting solvent-free lipid bilayers for more than 2 weeks, which was probably a world record. Independent of the exciting lipid work, this experience convinced me that microfabrication was a fabulous new tool for building sensors and for holding small volumes of fluids for, among other things, spectroscopy.

Immediately on arrival at UW, I began to use ‘tricks’ of optics from my previous work to devise sensors. My initial sensor work at UW was on two optical sensors for general anesthetics—first using the change in the phase state of lipids caused by dissolution in the lipids by the anesthetics. The second was based on monitoring the presence of anesthetics in a matrix of perdeuterated silicone rubber using Raman spectroscopy. In this latter project, we began to use lithographic techniques to pattern polymeric waveguides. Both projects proved academically publishable but not commercially viable.

By 1992 I had become convinced that collaboration on larger projects than single-investigator NIH or NSF projects was the route to significant sensor developments. With the help of my Chair (and now UW Acting President) Lee Huntsman, I organized a series of meetings to find points of common interest between several people on campus already working on sensors. We had, at UW, a great resource in the Washington Technology Center’s Microsensor Fabrication Laboratory. This gave us access to most of the silicon microfabrication techniques we thought we could use at the time. One idea that leapt to the fore came from Prof. Margaret Kenney from the Department of Laboratory Medicine. She was convinced, as the manager of one of the most centralized hospital stat-labs in the country, that point-of-care diagnostics were the wave of the future. It made sense that small instruments would require small parts, and so we were off. With help from Profs. Martin Afromowitz and Robert Kaiser, there ensued a series of failed attempts to raise funds for an initial project we called ‘The Portable Stat-Lab’. It was to use silicon microfabrication to create a laptop-sized instrument for monitoring vital blood chemistries (and hematocrit) needed in triage on the battlefield. We ultimately were about to give up on fundraising attempts when the phone rang—it was Dr. Robert Morff of a venture capital company called Senmed Medical Ventures. He had been a reviewer on one of our (failed) proposals; he proposed that his company support our work in return for the rights to the IP to be generated. Thereupon ensued 6 months of negotiations between Senmed and UW. Before they were closed, we went back to Ira Skurnick at DARPA, and by June 1994 we had a sizeable multi-investigator program funded from two sources on development of the Portable Stat-Lab.

By the time the project began in June 1994, we had included Kaiser, Afromowitz, Kenney, and Profs. Forster and van den Engh from Mechanical Engineering and Molecular Biotechnology, respectively. This multi-disciplinary team proved to be exactly what was needed to make rapid progress. What began as an idea to miniaturize existing macroscopic analytical processes rapidly became a process of mutual education of how things differed when the Reynolds numbers were low and the dimensions were small enough that diffusional transport dominated over convection. We had endless meetings, discussions, arguments, and moments of discovery. From these fundamental differences came a steady stream of novel ideas, papers and patents. We were just early enough in what was to become the microfluidics field that we were able to patent extremely simple things. It helped that while we understood microfabrication, and used it on a regular basis, we were not primarily fabricators. As it turns out, some of what we were ‘discovering’ was already known to those in Field Flow Fractionation, but we were just ignorant enough to be infatuated with our own discoveries. We were the one-eyed men in the land of the blind.

The key technologies from these early days were the separation/extraction device we called the H-filter (primarily from James Brody, now of UC Irvine), and the measurement device we called the T-sensor (primarily from Bernhard Weigl, now of Micronics, Inc.). These are described in some detail at my www site (http://faculty.washington.edu/yagerp). What made life easy was that these devices were both powerful, intrinsic to microfluidics, and preposterously simple to explain to anyone with a high school education. Since then our laboratory had been continuing to exploit these devices, in a series of publications and patents. The most recently issued patent is for a novel type of immunoassay based on the T-sensor, known as the diffusion immunoassay, of DIA. The DIA (largely due to Anson Hatch) is a competition immunoassay that relies on the change in movement of a labelled antigen across a microfluid channel depending on whether or not there are slowly diffusing binding sites present in a second flow stream. This assay is both sensitive, and extremely rapid, being complete in under 20 seconds. It is also compatible with microfluidic systems that would allow real-time calibration. These ideas are all compatible with practical point-of-care diagnostics. The successes in generating IP resulted in Senmed founding a company called Micronics in Redmond, WA, just across Lake Washington from UW. This company has licensed much of the UW IP from the early Senmed/DARPA project and subsequent work from my laboratory. Micronics also contributed another vital piece of the puzzle. Since their initial focus was on the use of disposable components as the revenue-generating stream for them, they found that silicon was too expensive to be included in those disposables. They found that they could fabricate H-filters, T-sensors and more using very inexpensive commercially-available laser cutting tools. The laser was used to cut out parts in Mylar, some of which were covered with pressure-sensitive adhesive, allowing rapid prototyping assembly of complete multilayer laminate microfluidic cartridges. This laminate format has been fully adopted in our lab for most of our microfluidic work.

A subsequent DARPA-funded project in the MicroFlumes program led us to explore the potential of the H-filter for sample preconditioning in the detection of chemical and biological warfare agents. We explored (and are continuing to explore) the use of electrophoresis and isoelectric focusing in microdevices, but unlike work in other laboratories, we applied the field across the flow stream(s). This was a kind of extension of the H-filter, and allowed development of useful continuous extraction and fractionation processes. We showed that as long as the potentials were kept below about 2.3 V, it was possible to operate bubble-free, and to generate stable pH gradients in the channel from a single input (weak) buffer. While these approaches will probably never replace conventional 2-D gel electrophoresis, for some tasks these methods are clearly superior. We are currently pushing the potential to higher levels by moving the electrodes outside the disposable laminate cartridge.

Today we focus on the integration of multiple microfluidic processes into complete systems, primarily for biomedical diagnostics. The aim is to create a disposable that can carry out perhaps 20 immunoassays in less than 5 minutes that can be stored indefinitely at ambient conditions, and can be read out in an instrument the size (and cost) of a cell phone or PDA. We are pursuing a set of methods for controlled delivery of dry reagents from storage depots on the laminate cards. In addition to optimizing the DIA for this purpose, we are also actively pursuing the use of surface plasmon resonance imaging as a platform to support such detection. The optical table that once held our Raman system now holds a simple scanning confocal microscope and a sophisticated but simple SPR microscope.

Conclusion

Microfluidics has now lived through the first cycle of excitement and commercial failures. The hype of the late 90’s has been replaced by unwarranted cynicism as to the field’s potential. The next wave will be based on real successes in products and markets that may differ substantially from those upon which the first wave focused. I personally feel that the future is in what we call distributed diagnosis and home health care (D₂H₂). This is based on the demographics in the US and abroad, a technologically overloaded healthcare system, the increasing threat from emerging diseases and bioterrorism, and the fact that we are learning so much about the processes underlying both health and disease. The technologies that best address these issues will be the first big successes in microfluidics outside of the ink-jet printer. With luck, this technology will be sufficiently inexpensive that it can be implemented outside of the wealthy countries of the world, where it is needed the most. I find myself able for the first time in my career to integrate my scientific and technological knowledge with my belief that we have both a duty and a vital interest to get better diagnostic technologies available wherever and whenever they are needed. This is very satisfying, and may finally convince my mother that not going to medical school out of high school wasn’t a complete mistake.

Prof. Yager’s Research Group

---

Footnote

† Electronic supplementary information (ESI) available: Text of interview with Professor Paul Yager. See http://www.rsc.org/suppdata/lc/b3/b304864k/

---