Slime mold and cancer cells "sniff" their way through mazes by breaking down chemical molecules

Research demonstrates that cells are capable of sensing, and creating, chemical gradients to travel long distances

Xinwen Zhu

Systems Biology

Boston University

Cells have an incredible ability to find each other over long distances. During embryonic development or the spread of cancer, cells travel through complex environments to migrate to precise locations, using chemical signals as guidance. Understanding how they use faraway chemical cues to travel long distances would help us grasp the underlying mechanisms of human development, inform cancer treatment, and even one day open the path to controlling cell migration for synthetic biology.

Chemoattractants are chemicals towards which cells have a tendency to move. Just as receptors in our noses sense chemicals in the air, receptor molecules on the surface of a cell can pick up chemoattractants diffusing from a source. Cells move in the direction from which they pick up the strongest whiff of chemoattractant – enabling them to respond to chemical gradients. This strategy works well when the chemical source is nearby. However, just as our smelling abilities are limited by range, gradients are much harder to detect at long distances. In extreme cases, the gradient could be so weak that it is imperceptible. How, then, do cells decide where to orient themselves? 

An emerging hypothesis is that cells manage these situations by breaking down chemoattractants in their immediate vicinity, thus generating sharp concentration gradients (differences in the concentration of the chemoattractant across space) near the cell which can be used to determine direction. For example, cancerous epithelial cells break down and generate gradients of chemoattractants to migrate from their primary tumors to other locations.

To investigate this phenomenon, a study published in Science in August of last year challenged slime mold and cancer cells to solve complex microscale mazes. The scientists constructed microfluidic mazes consisting of thin channels in a silicone mold bonded to a glass petri dish, using techniques similar to those currently used to build organ-on-a-chip models

The researchers placed cells at the start of the maze and evaluated their ability to make it to a large chemoattractant reservoir on the other end. The cells solved the mazes by sensing, moving toward, and degrading chemoattractants along the way. These results demonstrate that these simple cellular functions could be sufficient to account for how eukaryotic cells traverse long distances in complex environments.

The maze setup reveals how such self-generated local gradients can be used to navigate complex environments. In the team's experiment, the concentration of chemoattractant was uniform throughout the maze at the start, so any direction-indicating gradient had to be generated by the cells themselves. To prevent interference between cells during the maze runs, the researchers mutated the cells so that they could no longer produce the chemoattractant themselves, but they could still respond to it and break it down.

First up in the maze was the slime mold Dictyostelium discoideum. For most of its life, this amoeba is a single-celled organism, but it has a remarkable social behavior: when starved, thousands of slime mold cells come together over long distances to form multicellular structures. This aggregation is guided by a chemoattractant known as cyclic AMP, which the cells produce and break down in periodic pulses. 

Before attempting the full maze, the slime mold cells first ran a race where one group of cells was exposed to regular cyclic AMP and the other had a version that was chemically modified to no longer be degradable. Cells migrated faster when they were able to degrade cyclic AMP than when they were not, confirming the idea that cells are better at moving towards chemicals that they are able to break down. Breaking down the attractant directs cells to move away from areas of high cell density, because the amoebae deplete cyclic AMP faster where there are more cells. This effect encourages the cells to disperse from the starting area and keep moving forward. 

Notably, unlike another maze-solving slime mold, Physarum polycephalum, which approaches the task by exploring every possible path before picking the best one, Dictyostelium discoideum cells were able to avoid going down incorrect paths using their self-generated chemoattractant gradients. The researchers obtained similar results using pancreatic cancer cells isolated from mice, which also generated gradients by breaking down chemoattractant, suggesting that the principles explored in the study are not specific to just one cell type. 

Finally, the researchers also supported their findings by tricking the cells with what they called a chemoattractant "mirage," and they were able to correctly predict with computational simulations what types of mazes would be easier or harder for the cells to solve. For example, they predicted that dead ends that branched or widened and thus acted as reservoirs for chemoattractant would lead more cells astray than short dead ends that did not hold much chemoattractant. As expected, the cells easily avoided going down short dead ends but were more often misdirected by complex branching dead ends. By switching up the designs of the miniature mazes on their microfluidic chips, the researchers took full advantage of the artificial chip microenvironments to experimentally test the predictions of their computational model.

In addition to helping us understand how cells navigate complex environments to find faraway attractant sources, this research may be useful to synthetic biology efforts such as regenerative medicine and multicellular system design. To regrow a whole organ or to generate a new biological system, cells of different types must sort themselves out and use chemical cues to find their proper positions, just as they do during embryonic development. Existing tools allow us to engineer cells to be attracted to chemicals to which they are not naturally responsive. For long-range communication, perhaps it will be equally important to teach these cells to break down their new chemoattractants. 

Comment Peer Commentary

We ask other scientists from our Consortium to respond to articles with commentary from their expert perspective.

Simon Spichak


Great article Xinwen! I love the science of slime molds but never knew the similarities they shared with cancer cells. It also reminds me of the ways that brain cells travel and organize themselves over the course of development. The video looks incredible but I am curious as to why there’s always a group of cells that go the wrong way. Why are some of the more efficient at going through the maze than others? Is their movement initially random, and then corrected depending on where the chemoattractant gradient leads them?

Do you think this opens up new strategies for treating cancer cells and preventing them from spreading? A chemoattractant ‘mirage’ could confuse or control the growth of a cancerous tumor. Again a very fascinating piece and a great job breaking it down.

Xinwen Zhu responds:

Thank you Simon! I would agree with your suggestion that there is a stochastic element to slime mold migration - these amoebae are extremely motile and in the absence of chemoattractant cues they still wander quite a lot!

I’m not sure about the feasibility of altering chemoattractant concentrations in the human body to prevent cancer metastasis, but that is a very interesting idea! I personally think the main applications of the work are likely going to be in synthetic biology, but if a certain cancer was found to be guided by cues that could be easily manipulated this could be something to keep in mind.

Hazel Walker

Immunology and Cell Biology

University of Cambridge

This is a really great article Xinwen! The work also has interesting implications for our understanding of the immune response. Neutrophils are the first immune cells to reach sites of infection and injury and they also use chemoattractant gradients to navigate, often across large distances.

We know that the first neutrophils to reach an injury release chemoattractants to call for more neutrophils but less is known about how they might break down the chemoattractant gradient as they migrate along it and what effect this might have on their behaviour. Understanding how chemoattractant cells navigate in mazes takes us one step further to understanding how this might work in the more complex setting of the body!

Xinwen Zhu responds:

Thank you Hazel! You bring up good points about the immune response, and D. discoideum has been used as a model organism to understand neutrophil migration in the past. It would be really cool to follow up this work with immune cells.