Arnold Kriegstein is a developmental neurobiologist at the University of California San Francisco (UCSF; CA, USA). He is also a clinical neurologist and directs the stem cell program at UCSF, which is one of the largest stem cell programs in the USA. The program covers stem cells of all organ types and diseases; however, Kriegstein’s interest lies within the brain. He is particularly interested in neurodevelopmental disorders and brain development.
In this interview, Kriegstein tells us about his work on cerebral organoids, including how we might define a cell’s maturity, and if we could ever reach a point where we have one standardized definition of maturity for organoids. He also discusses some of the challenges involved in using organoids to study brain development and how these may be overcome.
You’re presenting a talk here at FENS (7–11 July, Berlin, Germany) on the use of cerebral organoids – could you tell us more about this and why you began to use organoids in your research?
I am especially interested in the development of the cerebral cortex of the brain. In humans, this is an area that is very distinct from any other living mammal. It’s actually very different compared to the mouse, and yet, the mouse is the model that most scientists use to study normal development of the cortex, and also diseases of the cortex such as autism or schizophrenia. I think it’s going to be very difficult to make the case that one is discovering the mechanism of human disease if you are using non-human models.
“We can, for example, grow organoids from patients with autism or schizophrenia, or a cortical malformation, and then study the mechanism of the disorder in the laboratory.”
My lab has been focusing on primary tissue samples that have been donated for observation. Those are very limited, of course, and are very precious. They are also not the type of tissue that we can experimentally manipulate, at least not very easily. We are turning to organoids as a way to investigate two things: first, to model human development in a reduced way but hopefully still more human-specific way than would be possible in an animal model; and second, to take advantage of the ability to derive stem cells and organoids from patients who carry disease mutations. We can, for example, grow organoids from patients with autism or schizophrenia, or a cortical malformation, and then study the mechanism of the disorder in the laboratory.
Some of the challenges involved with using organoids to study human brain development and disease include modeling advanced developmental stages or incorporating missing cell types. How close do you think we are to overcoming these challenges?
In studying developmental or disease mechanisms, one should always choose a reductionist model, and one that has enough features in common with whatever it is you are trying to study so that it can provide useful results. The same is true with organoids. I don’t think they are ever going to reproduce all of the features of the developing brain.
On the other hand, for most studies, one really wants a reduced model in order to eliminate some of the confounds, complexities and so on, of a really highly developed structure. It is probably sufficient if the model has the key elements that are relevant for the disease.
Currently, it is true that organoids are lacking certain critical cell types. Moreover, they don’t have certain brain regions, connections with those regions or with sensory organs. Thus, they are very isolated and restricted. However, they contain cell types that are of great interest. As long as those cell types develop in a reasonably normal context, one can compare them to disease cells and learn quite a bit about disease mechanisms.
Another question hanging over organoids is how do we define the cell’s maturity? Do they develop at the same rate as human cells? If the environment is different, can they classed as mature as the same age in a normal environment? Do you think we will reach a point where we have a standard definition of mature for organoids?
“These cells are at different stages of maturation because the progenitors or stem cells in the organoids are generating these cells over time… You have to somehow deconcolve the ages of the individual cells that make up the organoid. We have been doing that genetically.”
We have actually been tackling this problem over the last few years. We have been doing it on a single-cell level and what we find is that an organoid has a temporal age. They may have been in the laboratory for 1 month or more, but within that organoid, there are heterogenous populations of cells. These cells are at different stages of maturation because the progenitors or stem cells in the organoids are generating these cells over time. Some of these cells will be born early, and some will be born later. You have to somehow deconvolve the ages of the individual cells that make up the organoid.
We have been doing that genetically. We can dissociate the organoids, capture individual cells, and look at the gene expression patterns of those individual cells. This allows us to identify what cell type they are. Once we know the cell type, we can look at genes in those cell types that change as a function of age. That gives us a cell type-specific maturation map.
This has initially been done using donated primary tissue but not with organoids. We can take human tissue at different stages, disassociate it, do single-cell RNA sequencing, look at the gene expression patterns in the individual cell types, and find modules of gene expression that change over time, or with maturation, that are cell type-specific. If you just take all the genes in an organoid at once and look at them, you have what’s called ‘bulk gene sequencing’. You’ll never be able to disentangle which genes are related to which cell at what stage of maturation.
If gene sequencing is done on a single-cell level, you can then come up with networks of genes expressed by a single cell at a certain stage of its development. Once we know this in a normally developing brain, we can look for the same gene networks in the same cell types in organoids – and we have done that. The good news is that you can find the same network of genes that we think are signs of maturation in organoid cells as you find in primary tissue.
We have therefore been able to construct a maturation guide, or report card, for each individual cell type. We can calibrate our organoids according to their stages of maturation. That allows us to compare the same cell type across organoids in primary tissue or from one organoid to another; not only the same cell type, but the same type at the same stage of maturation.
“A real bonus has been that we can do this across species… we can find the same cell type and the same maturation genes that allow us to compare them to humans.”
A real bonus has been that we can do this across species. For example, in our organoids that come from non‑human primates, such as monkeys and chimpanzees, we can find the same cell type and the same maturation genes that allow us to compare them to humans. We now compare, for example, our favorite cells – the radial glial cells. A radial glial cell from a human can be compared with one from a chimpanzee with confidence, as they are at the same stage of maturation. Then we can focus on those other genes that they have in common, or that are differentially expressed.
So if we have the same cell type at the same stage of maturation, and the human cell has a set of genes that the monkey cell doesn’t have, then those genes could be human-specific and important for the function that makes that cell type different in humans than in monkeys. That’s really what we want to get at. We need to first calibrate all the other features in order to come up with those genes.
Where do you think we could be 10 years from now, and where might the best uses for these organoids be?
I think there is enormous excitement about organoids because they are human. There is almost no other way to study human cells, either during development or in a disease state in the laboratory. It’s something you couldn’t do in an animal model.
I think that for discovering cell types, gene expression networks, or disease-related molecular changes, this is going to be fantastic. To take the next step and figure out what the function of disease-related genes may be will probably involve animal models. It could be a mouse or it could be a non-human primate. People are considering the marmoset as an experimental model of a non-human primate. One can also use ferrets or other animals and use gene editing to study the functional effects of gene expression patterns.
To discover the genes that are important for human development, I think there’s no other way than by studying humans. Now, we can at least do that with organoids and human cells in the laboratory. That’s a huge advancement.
Another thing we can do is use organoids as drug-screening platforms, particularly if they express disease-bearing human cells. If we can find druggable targets or treatments that correct a defect in human cells in the laboratory, this makes it much more likely that it will actually work in human patients. There’s been a big problem with Big Pharma discovering drugs that work in mice but fail in human trials. This may be a way to get around that problem – to discover drugs that actually are effective in human cell models, organoids essentially, and then translate that into a treatment that might work in a patient.
You might also like:
The opinions expressed in this interview are those of the interviewee and do not necessarily reflect the views of Neuro Central or Future Science Group.