University of Pittsburgh

Machine Biology

Two married Pitt engineers, Anna Balazs and Steven Levitan, often discuss their research over dinner. One night, they were talking about how computers might be used to control the movements of tiny, nanoscale materials. Then, the conversation turned to ants.

Written by Jennifer Bails

An ant crawls stealthily across the kitchen counter, quickly descending into the sink in a hunt for food. The foray pays off with the discovery of a coffee teaspoon still dripping with cream and sugar. After a furtive taste of the sweet stuff, the ant speedily returns to its colony. Soon, fellow ants follow a scented trail into the sink, then return to the colony carrying a feast—sugar crystals.

Last summer, married Pitt engineers Anna Balazs and Steven Levitan saw this drama of nature unfold in their kitchen. “The ants came out of a socket and went all the way around the counter and into the sink to find the spoon of sugar,” Balazs recalls. “We actually sat there and watched it for a while—we were both fascinated.”

This insect encounter was riveting for the husband-and-wife team because it offered a real-life demonstration of what they had striven for months to accomplish in their research—the creation of “virtual” biology.

After ant-watching, the couple did what most homeowners do when summertime insects show up in the kitchen. Let’s just say it didn’t end well for the ants. But the ants’ behavior remained an inspiration.

For the past two centuries, scientists have been trying to understand the microscopic inner workings of cells and to make sense of the extraordinary complexity of these fundamental building blocks of life. Balazs, though, is not a cell biologist; she is a chemical engineer. Still, she draws on what is known about how living cells are made and how they function to inform the digitally driven, artificial systems she designs.

She seeks to create biomimetic materials—products that mimic processes or behaviors found in nature. In other words, biology is her inspiration for advances in engineering.

Since 2006, Balazs—Distinguished Professor of Chemical and Petroleum Engineering and the Robert Von der Luft Professor in the Swanson School of Engineering—has been working to build a system that mimics cell signaling—a vital biological process. Cells in organisms of all kinds have complicated machinery that allows them to share information with each other, often by releasing signaling molecules. These communication networks play a key role in cellular activities involved in everything from tissue repair to organ development to complete biological functioning.

It wasn’t Balazs’ goal to reconstruct the elaborate pathways that allow real cells to communicate with each other. Instead, she simply wondered whether she could devise a stripped-down version of a biological cell that could talk to other cells of its kind—without all the intricate parts used in actual cell signaling. “The real challenge was how do you mimic the complexity of biology when all you have are basic tools,” she says.

The basic tools she started with were microcapsules, which are fluid-filled balls enclosed by an elastic shell—no bigger than a micron in diameter. The period at the end of this sentence is roughly 400 microns. These cell-like microcapsules can secrete even more minuscule nanoparticles, much like the way biological cells release chemical signals to communicate with one another.

“Essentially we wanted to use purely synthetic materials to have one round thing send a message to another round thing, and through that communication, perform some function,” she says about her micron-scale experiments.

Postdoctoral researchers German Kolmakov and Victor Yashin helped Balazs develop a computer model of such a system using the high-powered supercomputers in Pitt’s Center for Molecular and Materials Simulations, which was established in 2000 to support computational research on campus.

In their model, the microcapsules exchange two different types of nanoparticles. A “signaling cell” begins talking by secreting nanoparticles known as agonists. These agonists then prompt a “target cell” to enter into the conversation by releasing nanoparticles known as antagonists. But the antagonists, in turn, stop the first cell from releasing agonists. That means once the signaling cell falls silent, so, too, does the target cell, which makes the signaling cell start talking again.

And so on and so on, the dialogue continues back and forth.

The most intriguing facet of Balazs’ model is that the nanoparticles released from a signaling cell make the surface underneath them stickier. This causes the target cell to move toward the signaling cell. But when the target cell gets too close, it is triggered to release antagonists, which make the surface less sticky, causing both cells to move away from that spot together.

In this way, groups of capsules begin to form as the signaling cell rolls along, picking up new target cells. And by changing the parameters of the system, the scientists discovered they could control the motion of these capsules to create different, creature-like shapes. One formation they have made resembles a two-headed dragon, with two cooperating signaling cells leading a long chain of target cells. They also have made curvy snakes, with competing signaling cells each pulling its own line of target cells.

In essence, the computer model was able to stimulate synthetic materials to act in a way that mimics basic communication between living cells. Virtual biology.

When Balazs was a youngster, biology was common fodder for conversation around the family dinner table. Her father had a degree in veterinary medicine and was passionate about his job, which involved toxicology research. “We talked about science and his work all of the time,” recalls Balazs. Her parents, along with thousands of other intellectuals, escaped Hungary following the 1956 revolution that was quashed by the Soviets. They first settled in Canada and later moved to New York.

Balazs earned her bachelor’s degree in physics from Bryn Mawr College in 1975, then completed her master’s and PhD degrees in materials science and engineering at the Massachusetts Institute of Technology. She joined the Pitt faculty in 1987 along with her husband, Steven Levitan, who is now the John A. Jurenko Professor of Electrical and Computer Engineering. He also holds a joint appointment in Pitt’s computer science department.

Levitan studies the design of new computing devices based on emerging nanotechnologies and is especially interested in how to build microchips through a process called bottom-up self-assembly, whereby tiny parts organize themselves into working circuits.

Just like in Balazs’ childhood home, science is a regular topic of mealtime discussion for the Pitt couple. “We usually talk about our work over dinner,” he says. “Often we just complain to each other,” he jokes, “but sometimes it is constructive.” Such was the case one late night a few years ago, during a conversation over a good bottle of wine, when Balazs was sharing her excitement about her progress on her microcapsule research. Levitan had just returned from a conference where his colleagues were discussing ant colony optimization—a hot new topic in computer science.

As Balazs and Levitan would later observe in their own kitchen, ants use pheromones to signal the most direct route to a food source. It’s a biological phenomenon called stigmergy in which social insects use their environment as a medium of communication rather than coordinating their behavior through any direct exchange of information.

The lone ant on their countertop laid down a trail of pheromones on its way back to the colony, “mapping” the direction to the sugar crystals. These chemical compounds alerted the other ants in the nest that there was a meal to be had. One by one, they followed the scent, reinforcing the shortest path to the sink with their own pheromones as they headed back to the colony so that other ants could also find their way to the food as quickly as possible.

Computer scientists have developed mathematical algorithms that aim to simulate this process as a way to pinpoint optimal solutions to difficult computational problems. Perhaps, reasoned Levitan, his wife’s microcapsules could behave like ants in this way, laying down a trail of “pheromones” for other cells to follow.

“After another glass of wine, I said ‘I’m going into the lab tomorrow to tell my team to get started on that problem,’” Balazs recalls, then laughs.

The researchers found that by delaying the nanoparticle release of one microcapsule group until another passes by and activates it, the trailing group would follow the chemical residue left by the leading group—like ants following a pheromone trail to food. “It was really this beautiful inspiration of Steven’s that brought all of this work together,” Balazs says.

Their results were published last July in Proceedings of the National Academy of Sciences, marking just the second time Balazs and Levitan had coauthored a study together in 26 years of marriage. Their paper made headlines in the scientific community and also caught the eye of techies worldwide, as videos of “ants,” “dragons,” and “snakes” went viral on YouTube.

Cool videos aside, the study marked the first time anyone had shown how to get inanimate, microscale objects to communicate with one another to carry out a concerted action, says Daniel A. Hammer, an international expert in cell behavior who is the Alfred G. and Meta A. Ennis Professor of Bioengineering at the University of Pennsylvania.

“What they have done that is so special is to show that you can basically make mimics of communicating organisms by using inert particles, without having anything that’s living whatsoever,” Hammer says.

The potential applications of these animal-like cells are limitless, according to Hammer. In biomedicine, they could act like trucks that haul drugs to cells inside the body. Or they could be used in microfluidics, which are small-scale devices used for rapid biological assays or to synthesize minute quantities of chemicals. Or they could be put to work carrying tiny electronic parts—or laying down the pattern for these components—in microchip assembly, which is what interests Levitan the most.

“It’s the question of how we control and predictably move material from one place to another on the microscale to do complex assembly without using little tweezers or electric fields or something else,” he says.

Meantime, Balazs has made further advances with her models, demonstrating together with postdoctoral researcher Amitabh Bhattacharya how chemical waves can be propagated along a string of microcapsules to mimic cell signaling over longer distances. Her team also has developed a snake that can pick up a bunch of nanoparticles and drop them off elsewhere, and they are working to vary the kind of information the snakes carry.

For Balazs, the real reward comes from seeing an amazing process like ant foraging behavior, developed over millions of years of evolution, come to life in her computer models in a way that could have far-reaching impact: “It’s quite stunning to me that if you put the right chemistry and physics together, you can mimic aspects of biology.”

-->