Self-steering particles go with the flow

Asymmetrical particles could make lab-on-a-chip diagnostic devices more efficient and portable.

Anne Trafton, MIT News Office

stretchy-electronics-4MIT chemical engineers have designed tiny particles that can “steer” themselves along preprogrammed trajectories and align themselves to flow through the center of a microchannel, making it possible to control the particles’ flow through microfluidic devices without applying any external forces.


self aligning nano 20131108111613-0

A slightly asymmetrical particle flows along the center of a microfluidic channel

Such particles could make it more feasible to design lab-on-a-chip devices, which hold potential as portable diagnostic devices for cancer and other diseases. These devices consist of microfluidic channels engraved on tiny chips, but current versions usually require a great deal of extra instrumentation attached to the chip, limiting their portability.

Much of that extra instrumentation is needed to keep the particles flowing single file through the center of the channel, where they can be analyzed. This can be done by applying a magnetic or electric field, or by flowing two streams of liquid along the outer edges of the channel, forcing the particles to stay in the center.

The new MIT approach, described in Nature Communications, requires no external forces and takes advantage of hydrodynamic principles that can be exploited simply by altering the shapes of the particles.

Lead authors of the paper are Burak Eral, an MIT postdoc, and William Uspal, who recently received a PhD in physics from MIT. Patrick Doyle, the Singapore Research Professor of Chemical Engineering at MIT, is the senior author of the paper.

Exploiting asymmetry

The work builds on previous research showing that when a particle is confined in a narrow channel, it has strong hydrodynamic interactions with both the confining walls and any neighboring particles. These interactions, which originate from how particles perturb the surrounding fluid, are powerful enough that they can be used to control the particles’ trajectory as they flow through the channel.

The MIT researchers realized that they could manipulate these interactions by altering the particles’ symmetry. Each of their particles is shaped like a dumbbell, but with a different-size disc at each end.

When these asymmetrical particles flow through a narrow channel, the larger disc encounters more resistance, or drag, forcing the particle to rotate until the larger disc is lagging behind. The asymmetrical particles stay in this slanted orientation as they flow.

Because of this slanted orientation, the particles not only move forward, in the direction of the flow, they also drift toward one side of the channel. As a particle approaches the wall, the perturbation it creates in the fluid is reflected back by the wall, just as waves in a pool reflect from its wall. This reflection forces the particle to flip its orientation and move toward the center of the channel.

Slightly asymmetrical particles will overshoot the center and move toward the other wall, then come back toward the center again until they gradually achieve a straight path. Very asymmetrical particles will approach the center without crossing it, but very slowly. But with just the right amount of asymmetry, a particle will move directly to the centerline in the shortest possible time.

“Now that we understand how the asymmetry plays a role, we can tune it to what we want. If you want to focus particles in a given position, you can achieve that by a fundamental understanding of these hydrodynamic interactions,” Eral says.

“The paper convincingly shown that shape matters, and swarms can be redirected provided that shapes are well designed,” says Patrick Tabeling, a professor at the École Supérieure de Physique et de Chimie Industrielles in Paris, who was not part of the research team. “The new and quite sophisticated mechanism … may open new routes for manipulating particles and cells in an elegant manner.”

Diagnosis by particles

In 2006, Doyle’s lab developed a way to create huge batches of identical particles made of hydrogel, a spongy polymer. To create these particles, each thinner than a human hair, the researchers shine ultraviolet light through a mask onto a stream of flowing building blocks, or oligomers. Wherever the light strikes, solid polymeric particles are formed in the shape of the mask, in a process called photopolymerization.

During this process, the researchers can also load a fluorescent probe such as an antibody at one end of the dumbbell. The other end is stamped with a barcode — a pattern of dots that reveals the particle’s target molecule.

This type of particle can be useful for diagnosing cancer and other diseases, following customization to detect proteins or DNA sequences in blood samples that can be signs of disease. Using a cytometer, scientists can read the fluorescent signal as the particles flow by in single file.

“Self-steering particles could lead to simplified flow scanners for point-of-care devices, and also provide a new toolkit from which one can develop other novel bioassays,” Doyle says.

The research was funded by the National Science Foundation, Novartis, and the Institute for Collaborative Biotechnologies through the U.S. Army Research Office.

Nanosensors Could Aid Drug Manufacturing

nanomanufacturing-2 CAMBRIDGE, Mass. MIT News Office: Chemical engineers find that arrays of carbon nanotubes can detect flaws in drugs and help improve production.      — MIT chemical engineers have discovered that arrays of billions of nanoscale sensors have unique properties that could help pharmaceutical companies produce drugs — especially those based on antibodies — more safely and efficiently.
Using these sensors, the researchers were able to characterize variations in the binding strength of antibody drugs, which hold promise for treating cancer and other diseases. They also used the sensors to monitor the structure of antibody molecules, including whether they contain a chain of sugars that interferes with proper function.
“This could help pharmaceutical companies figure out why certain drug formulations work better than others, and may help improve their effectiveness,” says Michael Strano, an MIT professor of chemical engineering and senior author of a recent paper describing the sensors in the journal ACS Nano.
The team also demonstrated how nanosensor arrays could be used to determine which cells in a population of genetically engineered, drug-producing cells are the most productive or desirable, Strano says. Lead author of the paper is Nigel Reuel, a graduate student in Strano’s lab. The labs of MIT faculty members Krystyn Van Vliet, Christopher Love and Dane Wittrup also contributed, along with scientists from Novartis.
Testing drug strength
Strano and other scientists have previously shown that tiny, nanometer-sized sensors, such as carbon nanotubes, offer a powerful way to detect minute quantities of a substance. Carbon nanotubes are 50,000 times thinner than a human hair, and they can bind to proteins that recognize a specific target molecule. When the target is present, it alters the fluorescent signal produced by the nanotube in a way that scientists can detect.
Some researchers are trying to exploit large arrays of nanosensors, such as carbon nanotubes or semiconducting nanowires, each customized for a different target molecule, to detect many different targets at once. In the new study, Strano and his colleagues wanted to explore unique properties that emerge from large arrays of sensors that all detect the same thing.
The first feature they discovered, through mathematical modeling and experimentation, is that uniform arrays can measure the distribution in binding strength of complex proteins such as antibodies. Antibodies are naturally occurring molecules that play a key role in the body’s ability to recognize and defend against foreign invaders. In recent years, scientists have been developing antibodies to treat disease, particularly cancer. When those antibodies bind to proteins found on cancer cells, they stimulate the body’s own immune system to attack the tumor.
For antibody drugs to be effective, they must strongly bind their target. However, the manufacturing process, which relies on nonhuman, engineered cells, does not always generate consistent, uniformly binding batches of antibodies.

Currently, drug companies use time-consuming and expensive analytical processes to test each batch and make sure it meets the regulatory standards for effectiveness. However, the new MIT sensor could make this process much faster, allowing researchers to not only better monitor and control production, but also to fine-tune the manufacturing process to generate a more consistent product.
“You could use the technology to reject batches, but ideally you’d want to use it in your upstream process development to better define culture conditions, so then you wouldn’t produce spurious lots,” Reuel says.
Measuring weak interactions
Another useful trait of such sensors is their ability to measure very weak binding interactions, which could also help with antibody drug manufacturing.
Antibodies are usually coated with long sugar chains through a process called glycosylation. These sugar chains are necessary for the drugs to be effective, but they are extremely hard to detect because they interact so weakly with other molecules. Drug-manufacturing organisms that synthesize antibodies are also programmed to add sugar chains, but the process is difficult to control and is strongly influenced by the cells’ environmental conditions, including temperature and acidity.
Without the appropriate glycosylation, antibodies delivered to a patient may elicit an unwanted immune response or be destroyed by the body’s cells, making them useless.
“This has been a problem for pharmaceutical companies and researchers alike, trying to measure glycosylated proteins by recognizing the carbohydrate chain,” Strano says. “What a nanosensor array can do is greatly expand the number of opportunities to detect rare binding events. You can measure what you would otherwise not be able to quantify with a single, larger sensor with the same sensitivity.” This tool could help researchers determine the optimal conditions for the correct degree of glycosylation to occur, making it easier to consistently produce effective drugs.
Mapping production
The third property the researchers discovered is the ability to map the production of a molecule of interest. “One of the things you would like to do is find strains of particular organisms that produce the therapeutic that you want,” Strano says. “There are lots of ways of doing this, but none of them are easy.”
The MIT team found that by growing the cells on a surface coated with an array of nanometer-sized sensors, they could detect the location of the most productive cells. In this study, they looked for an antibody produced by engineered human embryonic kidney cells, but the system could also be tailored to other proteins and organisms.
Once the most productive cells are identified, scientists look for genes that distinguish those cells from the less productive ones and engineer a new strain that is highly productive, Strano says.
The researchers have built a briefcase-sized prototype of their sensor that they plan to test with Novartis, which funded the research along with the National Science Foundation.
“Carbon nanotubes coupled to protein-binding entities are interesting for several areas of bio-manufacturing as they offer great potential for online monitoring of product levels and quality. Our collaboration has shown that carbon nanotube-based fluorescent sensors are applicable for such purposes, and I am eager to follow the maturation of this technology,” says Ramon Wahl, an author of the paper and a principal scientist at Novartis.