Bubble Bots: Simple Biocompatible Microrobots Autonomously Target Tumors

The potential of microrobots is enormous. These miniature objects can be designed to carry out actions within the body, such as sensing biomarkers, manipulating objects like blood clots, or delivering drug therapies to tumor sites. But working out how to make the tiny bots effective, biocompatible, and cost effective is challenging. Now a Caltech-led team has taken a huge step toward making the next generation of microrobots for drug delivery. They have simplified both the structure of the microrobots and their production method, while making the bots highly effective and "smart" enough to direct themselves to a tumor.

The team of Caltech and USC scientists describe the bubble bots and their successful application in treating bladder tumors in mice in a paper that appears in the February 2 issue of the journal Nature Nanotechnology.

The team, led by Wei Gao, professor of medical engineering at Caltech and a Heritage Medical Research Institute Investigator, previously used ultrasound imaging and magnetic guidance in an animal model to deliver miniature 3D-printed robots to a tumor where they could biodegrade and release their cargo: cancer fighting drugs. Those microrobots were fabricated in a cleanroom with specialized equipment and featured a hydrogel shell made from a jellylike polymer surrounding a microbubble. This shell helped propel the bots and provided excellent imaging contrast to allow researchers to keep track of them within the body.

"We thought, what if we make this even simpler, and just make the bubble itself a robot?" says Gao. "We can make bubbles easily and already know they are very biocompatible. And if you want to burst them, you can do so immediately."

The team developed a method for creating such simple bubble bots. Using an ultrasound probe, they agitated a solution consisting of BSA (bovine serum albumin, a standard animal protein often used in lab experiments) to make thousands of microbubbles with protein shells.

Next, the scientists took advantage of another feature of the protein shell, the abundant amine groups available on the surface. Amine groups are a collection of atoms featuring a carbon-nitrogen bond, that can easily be chemically modified. By binding to these amine groups, the researchers created two types of microrobots with different ways to control their movements. And anti-cancer drugs such as doxorubicin can successfully bind to the surface of both versions.

The scientists attached the enzyme urease to the surface of both versions of the bubble bots. Urease acts like a tiny engine to get the robots moving. The enzyme catalyzes a reaction with urea, an abundant waste product found throughout the body that serves as a kind of biofuel for the robots, yielding ammonia and carbon dioxide. Because urease is not uniformly distributed on the surface of the bubbles, over time, more of these products will build up on one side versus the other. That imbalance creates an asymmetric chemical environment around the bubble, generating a net "push" that propels the microrobots forward.

In the first version, the team attached magnetic nanoparticles to the surface of the bubble bots, making them magnetically responsive. With help from ultrasound imaging of the bots' interior microbubbles, the bubble bots could be steered with exterior magnets to head toward a target within the body.

But the researchers wanted to go a step further. "We wanted to make the robots more intelligent," Gao says. Knowing that tumors and inflammation produce high concentrations of hydrogen peroxide compared to normal cells, the team decided to bind an additional enzyme called catalase to the surface of a second version of the microrobots. Catalase drives a reaction with hydrogen peroxide, creating water and oxygen. Through what is known as chemotactic behavior, the catalase-bound bubbles automatically move toward higher concentrations of hydrogen peroxide, steering them toward tumors.

"In this case, you don't need any imaging; you don't need any external control. The robot is smart enough to find the tumor," Gao explains. "The bubble robot's autonomous motion, together with its ability to sense the hydrogen peroxide gradient leads to this targeting, which we call chemotactic tumor targeting."

Once the bubble bots arrive at their target, the scientists can apply focused ultrasound to burst the bubbles, releasing their therapeutic cargo. That strong bursting action enhances the drug's penetration into the tumor as compared to the slowly degrading hydrogel robots previously used by the team.

When the scientists injected mice with bubble bots to deliver anti-tumor therapeutics, they observed a roughly 60 percent decrease in the weight of bladder tumors over a span of 21 days, as compared to mice given the drug alone.

"This bubble robot platform is simple, but it integrates what you need for therapy: biocompatibility, controllable motion, imaging guidance, and an on-demand trigger that helps the drug penetrate deeper into the tumor. Our goal has always been to move microrobots closer to real clinical use, and this robotic design is a big step in that direction," says the paper's lead author, Songsong Tang, who completed the work during his time as a postdoctoral scholar in Gao's lab at Caltech; Tang is set to join the University of Science and Technology of China as a professor.

The paper is titled "Enzymatic microbubble robots." Additional Caltech authors are Hong Han (MS '23); Xiaotian Ma (MS '24); Payal N. Patel (BS '25); Ernesto Criado-Hidalgo; Jounghyun Yoo; Jiahong Li (MS '23); Gwangmook Kim; Shukun Yin; Di Wu (PhD '21); and Mikhail G. Shapiro, the Max Delbrück Professor of Chemical Engineering and Medical Engineering and a Howard Hughes Medical Institute Investigator. Authors from USC include Chen Gong, Junhang Zhang, and Qifa Zhou. The work was funded by the National Science Foundation and the Heritage Medical Research Institute.