A Danish–Swiss research group has systematically tested which signalling substances cause stem cells to develop into different types of nerve cells in small artificial “mini-brains”. The results give laboratories a far more reliable starting point for cultivating precisely the types of nerve cells they need for disease models – and, in the longer term, perhaps also for treatments in which new nerve cells can replace cells lost to diseases such as Parkinson’s.
From the outside, the brain looks like a soft grey lump. Inside lies an orderly swarm of nerve cells, each following its own programme for where it should be located, how it communicates, and how signals travel through the network.
Stem-cell researchers are trying to recreate this order in artificial mini-brains – but in practice this has proved far more difficult than first assumed. If researchers can get stem cells to form the right types of nerve cells at the right “addresses”, they can use the models to follow disease processes closely, test treatments and understand what goes wrong – without having to study these processes directly in a human brain.
The better the artificial mini-brains mimic the real thing, the more confident researchers can be that their results truly say something about patients.
That is why a Swiss research group at ETH Zurich in Basel, together with Danish researchers from the University of Copenhagen, has been trying to find a more systematic way to create the same order in artificial mini-brains. One participant is Agnete Kirkeby, Associate Professor from the Novo Nordisk Foundation Center for Stem Cell Medicine (reNEW), University of Copenhagen.
According to her, the idea behind the project is to cultivate small clusters of brain cells from human stem cells, expose them to many combinations of signalling substances and compile the results in an atlas showing which signalling programmes typically cause stem cells to develop into nerve cells from, for example, the cerebellum, hypothalamus or cortex – effectively giving researchers a map to guide how brain cells develop.
“Making a mini-brain is not difficult. Making nerve cells in the laboratory is not difficult. The difficult part is controlling the formation of the nerve cells so that they acquire a very specific identity and end up in the right places in the mini-brain rather than random places,” says Agnete Kirkeby. “This gives us a better idea of where to start when we want to make a specific type of nerve cell.”
Signal programmes give mini-brains their direction
Before this work, frustration with messy mini-brains was widespread in many laboratories. A typical process began with a clump of stem cells, which researchers could easily coax into developing into nerve cells using specific signalling substances. From there, things went off track.
Instead of forming a pattern resembling a real human brain, the clump became a mixture in which cells that would normally be far apart ended up side by side. In the same clump, cells resembling the forebrain could be seen alongside cells from deeper brain structures and a few that mostly resembled tissue completely different from brain tissue.
To avoid this chaos, Agnete Kirkeby and colleagues distributed small spheres of stem cells into tiny depressions in plastic trays, with each cavity acting as its own experiment.
Each depression has its own schedule for the chemical substances that mimic the brain’s own developmental signals – and together the experiments form a large map of how different signal programmes shape the brain’s cells.
In parallel, Agnete Kirkeby’s group in Copenhagen contributed to the study by growing the cells as continuous tissue under gradients of the same factors, rather than as separate spheres. This approach can better mimic the continuous sheet of tissue that forms the brain during fetal development.
Researchers test thousands of signal combinations
In some trays, the researchers add a specific signalling substance in a low concentration from day one; in others only later or at a higher concentration. In still others, they mix several substances that, in the fetal brain, particularly influence the cerebellum or deeper brain structures.
After three weeks, the researchers halt development and split the mini-brains into individual cells so they can see which genes are active in each cell. Each cell is given a molecular barcode that reveals the well from which it comes, and the researchers then read the gene activity using single-cell RNA sequencing.
They then match the patterns of gene expression with reference data from the early human brain so that each cell can be classified as, for example, a forebrain cell, hypothalamus cell or cerebellum cell, depending on the developmental programme it follows.
The researchers can thus both link the signalling programmes to specific cell types and check the results against independent data from human brain development.
Timing and dosage shape which nerve cells emerge
The study culminates in an atlas in which each signalling programme is linked to the mixture of nerve cells the researchers find after three weeks – effectively a map of how different signals guide the development of brain cells. The map is based on systematic signalling experiments, analysis of gene activity in thousands of individual cells and comparisons with reference data from the early human brain.
When researchers leaf through the atlas, two key factors stand out: the strength of the substances and the timing of their influence – two control knobs that largely determine which types of nerve cells emerge.
The results also show that adjusting one knob at a time is not enough. Researchers have to think in terms of entire sequences and combinations of signals that together control the cells’ developmental programme, she says.
She also notes that several substances influence each other. Two signalling substances can together open up a specific developmental pathway that neither can create on its own. In other programmes, one substance effectively acts as a brake on the other.
“We have long known that a very specific combination of growth factors at the right time and at the right dose is required to steer stem cells towards specialised subtypes of nerve cells from specific areas of the brain. This study shows that these combinations can now be tested on a completely new scale, making it possible to identify new ‘recipes’ for nerve cells faster than before,” she says.
Timing and dosage shape which nerve cells emerge
The study culminates in an atlas in which each signalling programme is linked to the mixture of nerve cells the researchers find after three weeks – effectively a map of how different signals guide the development of brain cells. The map is based on systematic signalling experiments, analysis of gene activity in thousands of individual cells and comparisons with reference data from the early human brain.
When researchers leaf through the atlas, two key factors stand out: the strength of the substances and the timing of their influence – two control knobs that largely determine which types of nerve cells emerge.
The results also show that adjusting one knob at a time is not enough. Researchers have to think in terms of entire sequences and combinations of signals that together control the cells’ developmental programme, she says.
She also notes that several substances influence each other. Two signalling substances can together open up a specific developmental pathway that neither can create on its own. In other programmes, one substance effectively acts as a brake on the other.
“We have long known that a very specific combination of growth factors at the right time and at the right dose is required to steer stem cells towards specialised subtypes of nerve cells from specific areas of the brain. This study shows that these combinations can now be tested on a completely new scale, making it possible to identify new ‘recipes’ for nerve cells faster than before,” she says.
From pear porridge to precise nerve cell types
Agnete Kirkeby and her research group are developing exactly these types of neurons in the laboratory for transplantation into patients’ brains.
Experience from previous studies suggests that the new cells may function in the brain for many years, because it takes a long time before the disease also affects them. Results from animal studies and the first early clinical studies point in the same direction, but remain limited.
“We have already transplanted eight patients with Parkinson’s disease as part of an ongoing clinical study,” she says.
She also sees the atlas as a tool that extends beyond Parkinson’s and can help researchers working on other brain diseases – because in practice it provides a map for controlling how specific types of nerve cells develop.
For example, a laboratory investigating diseases that affect the cerebellum can use the data to identify the programmes that produced the most cells with cerebellar identity and begin optimising there instead of testing random combinations.
“Researchers can use the atlas to see which combinations and concentrations produced cells resembling the cerebellum and use this as a starting point for new protocols, where they fine-tune their approach towards the specific subgroup of nerve cells,” she says.
The next round of mini-brain experiments
The atlas mainly covers the early stages, when the cells have just chosen their direction. Researchers still need longer studies where the mini-brains mature and the cells form circuits and activity that more closely resemble the brain’s rhythms.
“We have a much better starting point, but we still need to understand what the cells actually do in the longer term,” says Agnete Kirkeby.
When she looks at the dataset from the new study, she does not just see coloured fields in a table but a map of how researchers can control the development of brain cells much more precisely than before, rather than simply hoping that the right cell types will arise on their own.
She sees the next round of experiments in her own laboratory as an opportunity for patients in future studies to receive nerve cells that resemble their own a little more closely, so researchers can rely more on knowledge and less on luck when developing new treatments – and base those treatments more on evidence than on chance.
