Neurobots: Biological Robots with A Simple Nervous System

In 2020, researchers engineered living, programmable organisms from Xenopus laevis, calling them xenobots, which could self-assemble and even reproduce. Building on this work, scientists later developed anthrobots—self-assembling biological structures created from human cells. This progression naturally led to deeper questions about complexity and function within these biological robots or ‘biobots’.

At the Wyss Institute, neuroscientist Haleh Fotowat collaborated with developmental and synthetic biologist Michael Levin of Tufts University to explore what might happen if neurons were introduced into xenobots and what kinds of nervous system architectures could emerge. “All these questions really excited me,” said Fotowat.

Their findings, published in Advanced Science, introduced ‘neurobots,’ which are biobots enhanced with neuronal precursor cells.¹ These neurobots grew larger than typical biobots, had increased movement, and underwent substantial changes in gene expression. This work offers insights into how nervous systems self-organize and may lead to new advances in synthetic biology and regenerative medicine.

To build neurobots, Fotowat isolated ectodermal tissue from the animal cap of a Xenopus embryo. “That’s the part that in the future makes the frog skin…if you take that part out and let it grow in a dish, it would still develop cell types that comprise the skin of the tadpole,” she explained. The biobot developed various cell types, including multiciliated cells (MMCs) for movement, mucus-secreting goblet cells, ion-regulating ionocytes, and secretory cells.

She then incorporated neural precursor cells from another embryo before the tissue fully closed and let the constructs self-organize and develop over about a week. The researchers then investigated whether these neurons could migrate within the neurobots and potentially influence their behavior.

Using antibody labeling, they confirmed the presence of neurons and structures such as axons and dendrites. While all neurobots developed neural tissue, each formed a unique neural architecture. Calcium imaging further revealed that these neurons were electrically active, functioning as simple neural networks.

The addition of neurons also altered the biobots’ physical form and behavior. Neurobots grew larger and more elongated, and instead of simple linear or circular motion, some moved in more complex, spirograph-like patterns. Neurobots with higher levels of neural tissue also exhibited a lower density of MCCs, suggesting a possible interaction between neural development and cell-type distribution.

Aside from the physical and behavioral differences, the researchers investigated whether the addition of neurons also influenced gene expression. They assessed the transcriptome of three groups: neurobots, biobots, and sham neurobots. The biobots and sham neurobots were more similar to each other. Meanwhile, neurobots exhibited a profile with a significantly larger number of differentially expressed genes, many of which were related to nervous system development.

“It was really interesting to see that there is this overexpression of these genes that encode different stages of the visual processing, from the eye to the photoreceptors to the retinal cells,” said Fotowat.

Leonid Moroz, a neuroscientist at the University of Florida, who conducts research on ctenophores, or comb jellies, to study their nervous systems, and who was not involved in the study, agreed, sharing that this result surprised him. He described the work as “a good Pandora’s box,” in which this approach of using a simplified living machine to study biological systems opens the door to many new and unexpected questions, particularly about how these cells precisely coordinate and organize within the system.

He was curious to see how this approach worked at different developmental stages, how it might be complemented by broader comparative analyses in other invertebrate and vertebrate organisms, and whether it could help identify key players that contribute to—and potentially aid in treating—neurodegenerative diseases.

Building on these findings, Fotowat is now exploring new directions opened by this research. In particular, she aims to better standardize neurobots to more reliably assess emerging patterns and gain deeper insights into the fundamental biology underlying these systems.