Jul 26, 2019 6:00 AM

Scientists Can Finally Build Feedback Circuits in Cells

The circuitry could act as control mechanisms in safe, smart cells programmed to kill tumors and treat brain injuries.

White blood cell with circuits under microscope

If you buy something using links in our stories, we may earn a commission. This helps support our journalism. Learn more. Please also consider subscribing to WIRED

Hana El-Samad is a bit of a control freak. She’s been obsessed with the idea of injecting order into chaos since she was a little girl growing up in Lebanon during the height of the country’s protracted civil war. During the day, El-Samad would bury herself in the textbooks her mother, a math teacher, brought home—the unfailing law and order of numbers and equations her cloak against the uncertainty outside.

When she got to college, that predilection turned to a career. That’s when she found a field called feedback control theory—the study of how complex systems regulate themselves. “Everything just came into focus,” says El-Samad, now a systems biologist at UC San Francisco. “There is a system of checks and balances that make sure the craziest of us are put back in line, and that’s true in human societies, in ecosystems, and inside organisms. Sometimes those systems fail. And you get war. Or disease.”

She started building robots, designing feedback control algorithms to stabilize their movements. A few years later, as a PhD student in Ames, Iowa, she worked on automated flight control systems for Rockwell Collins, a US military aerospace contractor. But El-Samad worried that her work would wind up in war zones like the one she had grown up in. That’s when she switched to designing circuits for living cells instead.

She wasn’t alone. In the early 2000s, lots of electrical engineers and control theorists were joining the emerging field of synthetic biology, thinking they could design biological circuits to control the flow of information in a cell, just as the physical versions moved electrons around computer chips. They started designing strings of DNA never seen before in nature, and sticking these synthetic operating systems into bacteria and yeast. But these tools remained blunt. Sure, you could make a malaria-drug-belching E. coli, but you couldn’t tell it to stop making the medicine when there was too much, or to just make a certain amount every day at the same time. For that, you needed feedback control.

And that’s where El-Samad comes in—along with the ideas of a Russian-American engineer named Nicolas Minorsky. Every time you flick on your car’s calf-saving cruise control or bask in your Nest-enabled, cool but not cold home, you can thank Minorsky. In the 1920s, his penchant for calculus combined with patient observations of helmsmen steering US battleships allowed Minorisky to figure out the mathematical theory behind proportional-integral-derivative control, or PID. Today, PID algorithms are everywhere in the modern world; they run aircraft autopilot programs, keep manufacturing robots from crushing their human coworkers, and make smart thermostats, well, smart.

This week, El-Samad and a team of collaborators from David Baker's lab at the University of Washington announced that they had figured out how to make a biological equivalent of a PID algorithm. It starts with a designer protein—they call it LOCKR, short for Latching Orthogonal Cage Key pRoteins. Shaped like a cage with a latched door on one side, the LOCKR opens wide when it comes into contact with a preset molecule, revealing whatever functions researchers have hidden inside the cage.

For example, in one version, that function is a tag that condemns whatever is attached to it to a cellular garbage heap. Say you’ve got an enzyme that takes molecule A and breaks it down into molecules B and C. But you’re worried you’re going to wind up with too much of molecule C, which can be toxic. You design a LOCKR whose key is molecule C, and you fuse it to molecule A. As the enzyme makes more and more molecule C, it opens more of these LOCKR proteins, revealing a tag to send it, along with molecule A to the garbage dump. Without molecule A, the enzyme slows down its production of molecules B and C. And that, my friends, is a feedback control circuit.

Network a few LOCKR-bound molecules together, and you’ve got a circuit that can control a cell’s functions the same way a PID computer program automatically adjusts the pitch of a plane. With the right key, you can make cells glow or blow themselves apart. You can send things to the cell’s trash heap or zoom them to another cellular zip code. That’s what the scientists showed LOCKR could do in yeast. Their long-term goal is to use LOCKR and other small, synthetic molecules to program human cells to steer themselves to diseased tissues, including the hard-to-reach brain, and reliably drop off a precise payload of drugs.

Other synthetic biology pioneers were impressed with the advance. “This technology is quite cool,” says Tim Lu, a computational biologist at MIT and cofounder of Synlogic, a company that is reprogramming bacteria to fight cancer. It has the potential to offer faster feedback control than other approaches, he says. But there’s still a lot of work to do before you could start thinking about putting LOCKR into people: “One of the key things that needs to be evaluated going forward is potential immunogenicity.”

That’s on El-Samad’s to-do list—figuring out if LOCKRs trigger a person’s immune system before they can do what they’re designed to do. Her team will need to study those cells over many months, and even years, to see if they’re able to sense the covert control codes and rebel against them.

If they can, then human cells won’t be good at consistently delivering drugs, and that’s a major goal here. El-Samad has a large (she won’t say how big) contract from Darpa, the Pentagon’s moonshot division, to learn how to link multiple LOCKRs together in the hopes of treating traumatic brain injury—one of the most common injuries sustained by soldiers. El-Samad and Wendell Lim, a biophysical chemist at UCSF, are loading synthetic circuits into white blood cells they’ve engineered to target the brain. Such cells produce inflammatory and anti-inflammatory molecules; the trick is getting the mix just right. Not enough, and the brain can’t start to heal. Too much and they can kill neurons, leading to behavioral changes, compromised motor skills, and cognitive decline. The circuits El-Samad is designing will put the genes that produce those molecules under the control of LOCKR, to bring them into balance. Her team plans to begin testing in mice sometime in the next year.

It turns out you don't need rationally designed circuits to make microbes burp out biofuels, makeup, and medicines. You can just move over big chunks of DNA from one organism to another and call it good. So what if you lose a few yeast along the way? Live cell therapies changed all that, says El-Samad, referring to the growing number of FDA-approved cancer treatments that involve engineering human white blood cells to sniff out tumors. These treatments can be miraculous, but they also are unpredictable. Sometimes the T cells go overboard, secreting cytokine storms that have killed patients. Bringing these cells under more tight control has to be a priority if they’re going to be the future of medicine, she says.

“For the last 10 years, synthetic biology has been a field that was very exciting but really lacked a purpose,” adds El-Samad. Now the cells synthetic biologists make have to be “reliable and smart and rational. Because if patients die, that’s the end of it. So suddenly, I think synthetic biology has found a purpose again.”