Researchers studying early embryogenesis often use human embryonic (pluripotent) stem cells. But when they try to develop these cells into parts of the embryo in a petri dish, the researchers can never be sure that the cells have become what the researchers think the cells should become. A new computer model may help researchers to determine this.
Experiments with human embryonic cells are not very common since these cells and embryos are reserved for the most important experiments, and maximising their benefits is therefore important.
Researchers sometimes face challenges in confirming what the cells in their petri dish have evolved into.
Are specific cells part of the yolk sac? Are they part of the supporting structures (placenta) around the developing embryo, aspects of the embryo or something else? A new computer model can now provide the answers.
The model enables more precise research on individual cells in the earliest stages of embryogenesis, and this could have implications for drug discovery, improving artificial insemination and other advances.
“Sometimes we must ask research questions that can only be answered by experimenting with human embryonic cells. But being certain about what we are observing is crucial, including whether the cells in a petri dish are what we think they are. Our model can help to answer this,” explains Joshua M. Brickman, Professor, Novo Nordisk Foundation Center for Stem Cell Medicine, reNEW, University of Copenhagen, Denmark.
The research, which was carried by a PhD student and post-doctoral fellow Martin Proks and Nazmus Salehin from reNEW, has been published in Nature Methods.
Embryogenesis is unique
The earliest stages of embryogenesis are unique since stem cells can develop into all types of cells in and around the embryo. For example, studying the supporting structures that surround the embryo requires guiding the stem cells to develop in the direction of the yolk sac and the placenta.
Alternatively, studying various aspects of the embryo itself requires that the cells develop in that direction.
One problem is that the types of cells in the embryo develop rapidly in different directions and make relatively small changes that can lead to large differences later in the embryo or fetus, and these cells can look very similar.
This challenges researchers to confirm their identity based on appearance alone – which the researchers solved with their computer model.
Model encompasses an avalanche of data
The researchers collected all the available information about various embryonic cells from the scientific literature, focusing on data obtained through single-cell RNA sequencing, which helps researchers to investigate how different types of cells express their functions through RNA profiles.
RNA expression varies between cells despite having identical DNA, enabling the types of cells to be identified based on their RNA expression. Since the researchers incorporated all available information into their model, they can now determine the types of cells in new samples by analysing their RNA expression.
They can now perform single-cell RNA sequencing on their samples, feed the data into the model and identify the closest matching types of cells, essentially using all the data available from the precious human embryos donated for research.
“Today, single-cell RNA sequencing is very affordable and provides enough information to determine whether the cells in an experiment belong to one type or the other. This enables researchers to rapidly confirm whether they should continue the experiment or if they have created something else in the petri dish,” says Nazmus Salehin.
Exploring how drugs affect the earliest embryos
The model can do more than just identify known types of cells, also indicating how closely an unknown type of cell matches a known type.
For example, researchers can apply drugs to embryonic stem cells to study how this affects cell development.
The model not only provides a percentage match but also explains which parts of the genetic profile led to a certain result.
“This enables us to determine whether certain drugs affect the expression and development of the stem cells and whether the drugs can harm the embryo. Similarly, we can test substances to improve artificial insemination and determine how they affect the stem cells,” notes Joshua M. Brickman.
Using the model in research
The researchers note that the model can help colleagues working with specific embryonic cells. For example, some researchers study the cells that form the placenta and how this can improve fertility treatments.
These studies typically begin with embryonic stem cells in a petri dish, tracking their development into placental cells and examining how various drugs support this process.
“Our colleagues can use the model to determine whether they actually develop precursors to placental cells without changing the genetic expression in other cells,” explains Nazmus Salehin.
Similar models exist for mature cells
Martin Proks points out that the model addresses the missing preimplanation gap and expands used research tools with interpretability/explainability.
Large databases of cells from adult humans already contain data on all types of cells in a specific organ and can be used to confirm the identity of the cells in experiments, but until now tools like this were not available for embryonic cells.
“Research on embryonic cells faces two major challenges: limited access to human embryonic cells and difficulty in distinguishing between closely related types of cells in early embryogenesis. This model solves both problems,” says Martin Proks.
The model can answer many questions
Joshua M. Brickman envisions broadly applying the model in human embryonic cell research. Current cell models for drug testing often claim to represent fetal cells but may not accurately mimic their properties.
The model can verify these claims and provide insight into congenital genetic disorders and how they affect early embryogenesis.
The researchers have also developed a similar model for mouse embryos, since mice are often used in preclinical studies. Comparing fetal development across species can also answer intriguing evolutionary questions.
“For example, humans and other mammals are genetically very similar but look different because of early embryogenesis. Comparing these processes across species may elucidate when the differences between humans and other primates begin to emerge,” concludes Joshua M. Brickman.
