By Diagnostics World Staff
October 5, 2015 | This week, a paper appeared in the Proceedings of the National Academy of Sciences demonstrating a device that could distinguish between three different strains of the influenza virus in as little as two minutes. Those results are comparable to the kind of diagnostic test that might be run in a hospital lab ― but this experiment was performed on a silicon chip no bigger than your thumbnail.
Senior author Holger Schmidt, whose lab at UC Santa Cruz designed the device, is a specialist in optofluidics, a field that involves directing light through tiny liquid channels. Although a physicist and electrical engineer by training, Schmidt now devotes a great deal of his time to the problem of making point-of-care diagnostics: miniaturized tests that can be run in remote locations with little in the way of lab equipment. In this emerging field, tools he originally developed with Aaron Hawkins of Brigham Young University to solve problems in atomic physics have opened up new ways to analyze clinical samples. The result is a prototype device that can identify single proteins or nucleic acids without the use of microscopes, providing an alternative to standard tests like PCR (polymerase chain reaction), in which bulky instruments must be used to make many copies of a DNA or RNA sample before it can be analyzed.
“Our vision was that this instrument would be particularly helpful for situations like with Ebola,” Schmidt tells Diagnostics World, “where you go to low-resource settings with people who may not be well trained, and conditions are difficult.”
Schmidt and Hawkins’ major contribution to optofluidics involves a type of structure called an antiresonant reflecting optical waveguide, or ARROW, which physicists have been designing in various forms since the 1980s. ARROWs use chambers lined with highly reflective materials to direct light down controlled paths. In the early 2000s, Schmidt and Hawkins helped develop new, hollow ARROWs etched into miniaturized chips, which could be used to guide light through minute volumes of liquids or gases.
The chips were originally designed for atomic spectroscopy, a method used in physics to identify atoms by the light they emit. (“We actually did make the first integrated atomic spectroscopy chip a couple of years ago,” Schmidt notes.) But the tiny devices also suggested one possible solution to a key problem in point-of-care diagnostics. Advances in microfluidics, controlling the flow of liquids through microscopic chambers, have allowed scientists to recreate many key lab processes inside chips, but reliably measuring test results in these microenvironments is a special challenge. Optofluidics provides a promising answer.
Single Molecule Detection
Many lab tests use fluorescent labels to provide their signals. In these tests, a molecular stain is bound to an analyte of interest ― say, a piece of viral RNA, or a protein on a virus’ coat ― and if the sample lights up, you know you’ve found a match. With an optofluidic chip, this process can be applied on a much smaller scale.
The key is overlapping a liquid channel, where the test sample passes through, with a solid-core waveguide to excite the molecular stains. “Each particle is like a little flash lamp, and when it passes by the optical excitation spot, it flashes and lights up,” says Schmidt. The flash is then carried directly to a light-sensitive detector, making the device sensitive to even a single fluorescently labeled molecule.
An advantage of this method is that the same chip can potentially be used for hundreds of different tests. “What’s lighting up is the stain,” says Schmidt, not the specific RNA sequence or protein you’re testing for. Anything you can stain with a fluorescent label, you can run through the chip.
The UC Santa Cruz lab has been experimenting with different procedures to apply those labels to clinically relevant samples. Two weeks ago, a paper in Nature Scientific Reports showed that their chips could be used to detect genetic material from the Ebolavirus, with very high sensitivity. In that study, a stain called SYBR Gold, which binds directly to nucleic acids, was fixed to Ebola RNA before the mixture was transferred to the chip.
This was the first time Schmidt and Hawkins’ chips were shown to pick up signals from clinically relevant samples (in this case, provided by the Texas Biomedical Research Institute, which has the secure facilities needed to work with a biohazard like Ebolavirus). “What was important to us in this case was to demonstrate that we can really see single molecules without expensive microscopy,” says Schmidt.
Notably, the sample preparation methods used in that study could be applied to any type of DNA or RNA analyte. The team used a batch of magnetic microbeads bound to fragments of nucleic acid that matched an RNA sequence unique to the Ebolavirus. When mixed with a sample of Ebola RNA, the microbeads grabbed the RNA from the solution and held tight to it. Then a magnet held the microbeads in place while everything else from the solution was washed away. As a result, by the time the fluorescent stain was applied, only the targeted Ebola RNA remained.
Simply substituting different microbeads would let users choose different genetic targets for the same assay. “It’s not only for infectious diseases,” says Schmidt. “We’re also working on cancer biomarkers ― you can look for genetic sequences that are specific to cancer.”
The results were encouraging: after testing samples at concentrations as low as 2.1 viral particles per milliliter ― consistent with the lowest concentrations detectable by, for instance, PCR testing ― the device was able to pick up a fluorescent signal within ten minutes. The team also tested samples of the closely related Sudan virus and Lake Victoria Marburg virus, showing that the sample preparation process successfully filtered out these control samples.
“For all these other viruses, no matter how high a concentration we use, we never got a single count,” Schmidt says.
One ARROW, Many Targets
The Ebolavirus paper demonstrated a nearly complete prototype diagnostic. Through a collaboration with Richard Mathies of UC Berkeley, the team was able to not only detect viral RNA on their optofluidic chip, but also perform a lot of the sample preparation on microfluidics ― although they stopped a little short of completing a front-to-back test that could be run outside a lab.
Schmidt says that such a test is still in the works, and that the obstacles are now more in the realm of engineering than proving the science. But in the meantime, his team has ambitions to expand their technology beyond looking at one viral target at a time. “If you’re looking for influenza and you send your sample to Quest Labs, for instance, they usually test for eight or nine different things at once,” he says. “So we came up with a new way to do that on our chips.”
On many lab-scale instruments, it’s possible to “multiplex” tests by using multiple fluorescent labels, all engineered to bind to different targets. That way, by detecting which labels successfully fix themselves to a sample, you can test for several different viruses at once.
Unfortunately, that process usually requires multiple detectors, each set up to monitor only one of the fluorescent labels. On a tiny optofluidic chip, this system of multiple detectors and split signals would be difficult and expensive to implement.
In the PNAS paper published today, Schmidt and his colleagues demonstrated a different way of designing multiplexed tests on their chips, choosing three strains of the influenza virus as their targets. The three strains were given separate fluorescent labels as usual. (Unlike with the Ebola test, in this case the labels were fixed to the viruses directly instead of to their RNA.) But when these fluorescent stains were excited inside the chip, all of their light signals were conducted by the ARROW to the same detector, rather than being split apart.
To figure out which label was exciting the detector at any given time, Schmidt and his team took advantage of the properties of different colored light. They replaced the simple solid-core waveguide from the chips used in their Ebola assay with a structure called a multi-mode interference (MMI) waveguide, which creates well-defined spots where fluorescent molecules can be excited. However, exactly where those spots fall and how many there are depends on the wavelength of light the fluorescent molecules are emitting.
That means a blue stain, with a short wavelength, will “flash” in a different pattern than a red stain, with a long wavelength, as they pass through the MMI waveguide. As the ARROW conducts these light signals to the detector, the device can simply count how quickly the stains are being excited: a fluorescent label more in the blue range will create a more rapid pattern of signals.
“Because the color information is in the time-dependence of the signal,” says Schmidt, “we don’t need to split up the colors and can use a single detector. All de-multiplexing is done by signal processing ― software ― which is much preferable and cheaper.”
The method is not infinitely flexible. Only a limited number of fluorescent stains can be used together, before their excitation patterns start overlapping and become impossible to tell apart. But by combining stains ― for instance, labeling one strain of flu with a red dye, a second with a blue dye, and a third with both dyes together ― the UC Santa Cruz team believes it should be possible to test for as many as 15 targets at once. That would be a major advance for point-of-care testing, letting users tell apart different influenza strains, or consider several cancer mutations together, or perhaps even test for bacterial infections and antibiotic resistance mutations at the same time.
“It is important to develop a versatile diagnostic instrument that can be used to analyze various biomarkers and bioparticles,” says Nastaran Hashemi, a specialist in point-of-care diagnostics at Iowa State University who was not involved in this research. “To expand the available spectral bandwidth in a specific photonic structure, using a single multi-mode interference waveguide, is novel.”
New Prototypes
Schmidt, Hawkins, and their colleagues still have a long way to go before their optofluidic chips could ever be deployed for live clinical testing. More sample preparation steps need to be transferred to microfluidics before the device will make up a complete point-of-care diagnostic ― especially in the multiplexed version of the test, where several different fluorescent dyes have to be applied to separate analytes.
The team has also never run a test starting with a raw blood sample, as would be needed in a doctor’s office or field clinic. In part, that’s because their first target was Ebolavirus, making blood samples extreme biohazards that can only be studied in specialized facilities, like the Texas Biomedical Research Institute that cooperated on the Nature Scientific Reports paper.
“What we’re working on right now is to build a prototype that they can actually employ in San Antonio in their facility,” says Schmidt, “until we can start with raw blood from infected primates, or whatever samples they have on hand.”
Most of all, the group needs to show that their test is highly reproducible, running tests on a large volume of clinically relevant samples and getting consistent and accurate results. Most proposed diagnostics never make it past this hurdle, especially in the point-of-care field, where few technologies are well-established.
Nonetheless, Schmidt is hopeful that his chips can one day be commercialized. He has already been a co-founder of one UC Santa Cruz spinoff company, LiquiLume, which aimed to produce optofluidics-based tests before it shut down in 2013.
“LiquiLume was a little bit ahead of its time,” he acknowledges, but adds that “of course we’re thinking about launching another company with these targets, because we’ve made tremendous progress.”
For now, his idiosyncratic team will continue applying physics principles to problems in the life sciences. Schmidt’s lab draws on the same unusual field of expertise that led him from atomic physics to diagnostics himself. Hong Cai and Joshua Parks, lead authors on the Nature Scientific Reports paper, are an engineering post-doc and a PhD student with a degree in chemistry, respectively, while lead author Damla Ozcelik of the PNAS paper is another engineering student.
“I think that’s very indicative of the kind of group we have,” says Schmidt. “We have everything from physicists to bioengineers and chemists.” And as the future of miniaturized diagnostics approaches the single-molecule scale, that’s a useful field of knowledge to be able to draw on.