Join MySLAS Social

Lab-on-a-Chip Technology: Collaborating to Confront the Chasm between Research and Product Development

Fostering collaborations between academia and industry for life sciences R&D is a key goal of SLAS. One area in which such collaboration is sorely needed is microfluidics, according to SLAS member George Whitesides, Ph.D., Woodford L. and Ann A. Flowers University Professor of Chemistry at Harvard University. In a recently published editorial, Whitesides underscores the "disconnect" between academic research on lab-on-a-chip technology and what users need in terms of products. In a subsequent interview, he expands on this theme, and other SLAS members share their views.


"There's a bit of confusion in the academic world," Whitesides says. "The belief is that making something that's easy to use, gives good results and costs very little is just development work, whereas science is about making things that are very sophisticated and complicated. But it turns out that it's actually quite difficult to make something that solves a real problem easily."

Academics frequently get involved in exploring "elegantly imagined" and "stimulating" devices, Whitesides observes. However, those devices often don't solve any problems in the real world, he contends. Users, on the other hand, generally are not focused on the intricacies of technology; they're interested in solving problems "simply and inexpensively," and as far as they're concerned, the technology should be invisible.

"These are different motivations," Whitesides says. "But researchers need to think about the difference between a measurement in a scientific paper and a product that is a solution to a problem. Both are interesting and probably required for the progress of science. However, if you make lots of things that make really great papers, but nobody ever uses them, it's a bit sterile." As an example, he points to work on open-channel PDMS (polydimethylsiloxane) microfluidics, "which has produced a flood of absolutely first-rate fluid physics and demonstrations of new kinds of things you can do on smaller samples. The problem is, they aren't being used very much." Although many academic scientists had expected microfluidics to be widely used in the laboratory by now, "the devices are still a bit too complicated and most are not commercially available."

Does New Mean Better?

Among the obstacles to wider use and development of lab-on-a-chip technology is the fact that many microfluidics applications—for enzymatic analysis, for example—already can be done on the "macro" scale, and researchers haven't done all the work needed to show that results achieved on the "micro" scale are comparable, according to SLAS member Gary Kramer, Ph.D., a research chemist at the National Institute of Standards and Technology. "On the micro side, researchers tend to give reproducibility information from their studies—'we did such and such on our microfluidic device X times, and here's the standard deviation of our results, and so forth.' It's a precision statement, but there's no accuracy statement to accompany it," Kramer says.

As an example, Kramer cites a study done by a biotech company years ago that he says still has relevance today. To check the results of a specific enzymatic assay, the company set up one robotic system to do the analysis and one microfluidic device to do the same analysis. "Both analytical schemes gave highly reproducible results; you could run the analyses again and again and again and get tight precision," Kramer says. "But the conclusions that could be drawn from looking at the data didn't match up at all—one of them was either right or wrong." Kramer believes the two systems were measuring different things, given the differences in parameters such as volume and surface area. "When you go to the micro scale, surface area is greatly enhanced over volume, and it's possible this created very real differences in physics and chemistry between the two technologies."

Such differences are not unprecedented, according to Kramer. In medical research, similar disparities occur between in vitro and in vivo tests. "Many people have tried to rationalize these differences," he says. "But parameters such as surface-to-volume ratio can explain why, in both settings, when people are trying to measure the same things, they're actually measuring different things because of the technology they're using. For microfluidics to live up to the hype people have built into it, researchers must prove that their new technique either gives related results, or similar results, to existing techniques, or they've got to be more specific about what they're actually measuring. Those measurements need to correlate with end results such as patient outcomes. This has started to happen, but very slowly."

More Time Needed?

SLAS2015 Conference Chair Elliot Hui, Ph.D., assistant professor of biomedical engineering at the University of California, Irvine (UC Irvine), agrees, stressing that the technology "is taking longer than people hoped" to make its way into the clinic. This is similar to what happened in a related field, microelectromechanical systems (MEMS), he says. MEMS started in the late 1980s as an effort to miniaturize mechanical structures onto silicon chips in a similar fashion to electronic circuits, Hui explains. That work was funded largely by the U.S. Department of Defense, which was interested in building devices such as small navigational systems.

By the early 2000s, MEMS was in the same situation George Whitesides describes with respect to microfluidics—people started complaining that there was a great deal of university research, but that work wasn't being deployed.

"Now, 10 years later, MEMS is everywhere—in cell phones, in Wii games and many other applications. It's a huge industry, but it just took longer than people anticipated," Hui continues. "I wouldn't be surprised if microfluidics followed a similar path, with a large investment of research at the university level, and at some point, a jump from there to the market."

The other point to keep in mind is that although microfluidics may not have made a big mark in the life sciences as yet, "if you think more broadly, the ink jet printer head is a microfluidic device," Hui notes. "It showed the value of the technology way before people started doing the biology, and it became a huge market. So I wouldn't count out microfluidics yet."

Industry Conundrums

In a similar vein to Hui, SLAS member and Buckeye Growth Partners founder and managing director Kevin Hrusovsky, member of the Board of Directors of several disruptive technology companies, including 908 Devices, argues that there have been major advances in microfluidics over the past five years, "but the learning curve probably took five to 10 years longer and significantly more investment from industry than was anticipated," he says. During that time, "some investors got burned and were turned off by the technology, and reigniting interest has required a reconstruction of mindsets around the potential of microfluidics."

When Hrusovsky took the helm at Caliper in 2003, the company's enterprise valuation had gone from $2 billion to negative $70 million, largely because of its commitment to microfluidics technology, he recalls. "The reality is, the ability to produce a commercial application is fraught with many challenges when you're dealing with this level of microfluidics. These include the quality of the chips, the reproducibility of the chips and the ability for the chips to add functionally to existing technologies.

"There was a point in time when Caliper's valuation imploded on its inability to deliver against investor expectations of the microfluidics IP [intellectual property] that came out of academia and/or government," Hrusovsky continues. "That implosion probably was not atypical for a pioneering technology. But I think in the past five years, Caliper and other companies have been able to demonstrate repeatable platforms, significant revenue and significant buy-in from users. We definitely are on the right path."

Academic Conundrums

"Some people believe it's not the job of the university or university-based research to do engineering development tasks or even to think at the beginning about the cheap, simple way of doing things," Whitesides observes. "They believe the job of the university is to demonstrate just the intellectual point that it is possible to make a measurement of a particular kind, not necessarily to build into it the idea that somebody else might eventually use it. But in a sense, if nobody uses it, it might as well not have happened." That said, he acknowledges that no one would have guessed at the beginning that the first NMR spectrometer would evolve into today's MRI imager, for example. "So, sometimes you have to demonstrate the phenomenon and then other people go to work to make it simpler, better, faster and cheaper."

Another conundrum involves funding. Although companies are investing more in microfluidics, funding agencies such as the National Science Foundation continue to emphasize basic research rather than practical applications, according to Hui. However, he notes that has started to change with the advent of small business grants and funding from philanthropic organizations that are focused on the development of devices that can be used in less developed countries, which provide opportunities to focus on practical applications.

Another issue is journal publication. "As George Whitesides points out, simple, cheap devices might not be published in the literature as easily as basic science," Hui says. "But here again, things are changing. Several years ago the Gates Foundation had a specific call for cheap microfluidics for diagnostics and several labs received money to develop these devices. Now a number of papers about such devices are starting to appear." Watch for the Journal of Laboratory Automation two-part special issue on Advancements in Biomedical Micro/Nano Tools and Technology, publishing in December 2013 and February 2014.

Collaborations, with a Caveat

Collaborative efforts between universities and industry have emerged to help resolve some of these conundrums and bridge the gap between microfluidics researchers and users. Hui is part of the recently formed National Science Foundation Industry/University Cooperative Research Center (I/UCRC) initiative called the Center for Advanced Design and Manufacturing of Integrated Microfluidics (CADMIM). The center's mission is "to develop design tools and manufacturing technologies for integrated microfluidics targeting cost-effective, quick and easy assessment of the environment, agriculture and human health." The academic researchers come primarily from UC Irvine and the University of Cincinnati, Hui says. However, the I/UCRC is open to any number of companies (about 20 have expressed interest so far), and Hui invites all SLAS industry members to get involved.

The upcoming I/UCRC is being developed on the heels of a similar program, Micro/Nano Fluidics Fundamentals Focus Center (M3), which was funded by the Defense Advanced Research Projects Agency (DARPA) and recently closed. "That center was made up of about 10 universities, but we had only a small number of companies involved," says Hui, whose group was part of that program, as well. An avid proponent of this collaborative model, Hui was a member of a similar program as a graduate student working in the MEMS field. The Berkeley I/UCRC "played a key role in launching the MEMS industry," he says, "and I have similar hopes for CADMIM."

While such collaborations will undoubtedly be helpful, Hrusovsky cautions that it's important to maintain a balance. "If academics become too enamored with making a commercial application part of their research, they may actually limit their potential to create breakthroughs," he says. "Although the pendulum can be swung a bit toward application know-how and competence, we don't want to lose the 'big-sky' discovery that the commercial people could then follow up on. In addition, venture capitalists who 'get' the value of 'disruptive innovation' understand the delicate balance between an academic's scientific capability and the ability to commercialize an application, which may take five to 10 years." A disruptive innovation, according to Hrusovsky, is one that helps create a new market and value network, and goes on, over time, to disrupt an existing market and value network, displacing an earlier technology. "Those venture capitalists are willing to wait for IP from academia to bear fruit. By contrast, investors who are in it for a short-term return may end up discounting and slowing down the potential of these new technologies."

A Bigger Picture

"The chasm between invention and a product is really about the 'D' in 'R&D.' That's not only a problem with microfluidics; that's a national problem for us in the United States," says NIST's Kramer. "As a nation, we're pretty good at inventing things, but we're not as good at bringing them to fruition as products. Unless there's a real market driver for something, it rarely gets funded here. Therefore, many newer technologies—first it was lab automation with robots, then microfluidics and now the same thing is happening in nanotech—have been way overhyped because the hype can lead to funding." By contrast, "countries that are willing to invest in the 'D' part, such as Japan, China, Korea and Germany, are moving full steam ahead."

To bridge the chasm in the United States without resorting to hype, Kramer advocates "getting graduate students and post docs out of their cozy research labs and into the real world, no matter what they're researching. Spending three months in a clinical lab would go a long way toward helping them understand what goes on there, especially in the area of quality assurance," he says. "They would realize that something that works in their own particular batch of cells might not necessarily work across all of mankind."

September 16, 2013