Ways of differentiation

Stem cells can differentiate into the final state in different ways

Polina Loseva, "Elements"

Fig. 1. Motor (motor) neurons derived from stem cells. Blue – DAPI, cell nuclei. Green – proteins confirming that these neurons are functional: Tubb3 – tubulin, a protein of microtubules in axons; Map2 – a protein associated with microtubules; VACht – a protein transporting vacuoles with the neurotransmitter acetylcholine. An image from the article being discussed in eLife.

Differentiation of stem cells, that is, their transformation into one or another cell type, is a complex multi–stage process. On the way to the final state, the cell goes through a number of intermediate stages. There are two approaches to cell differentiation in vitro. It is possible to consistently reproduce the processes occurring in the embryo during development, and gradually lead the cell in the desired direction. And you can immediately express proteins characteristic of the final state in the cell. Will the cells go through all the intermediate states in this case, or will they miss some? Or will they go the other way altogether?

The development of the embryo from a zygote to a full-fledged organism can be represented as a road with many forks. As cells divide, each of them chooses the path along which it will move on, that is, it acquires some characteristic properties (shape, internal structure and expression of specific genes). As a result of many such successive "decisions", the cell reaches the final state – one of the cell types of an adult organism with its entire set of characteristic features. This whole path of the cell as a whole is called differentiation, and the result is a differentiated state.

In 1954 , an English biologist Conrad Waddington proposed a model of the epigenetic landscape, which, in particular, is applicable to the differentiation of stem cells. This landscape looks like a slide with parallel branching ruts (creodes), and the cell appears to be a ball rolling down this slide (Fig. 2.). According to this model, the fate of the cell is determined once and for all, and it will not be possible to change it. In order for a cell to be in another organ instead of one, it needs to jump into a neighboring track, which is energetically unprofitable in this model.

Fig. 2. The epigenetic landscape of Waddington. Under each track, the final purpose of the cell is indicated: chest, wing, upper leg, lower leg, antenna, mouth (we are talking about the development of drosophila). Image from the book: C. H. Waddington. "Principles of embryology". NY, 1956.

Waddington's ideas remain relevant to this day, but experiments with stem cells suggest new rules for dealing with this landscape. So, in 2006, Japanese scientists Takahashi and Yamanaka managed to "throw a stone back on the mountain" – to return mouse cells from the terminally differentiated state to the level of embryonic stem cells of the embryo, from which any cell type can be obtained (see K. Takahashi, S. Yamanaka, 2006. Induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures by defined factors). To do this, 4 key transcription factors providing the stem state were injected into mature mouse fibroblasts. This process was called cell reprogramming. However , Yamanaka himself noted the low efficiency of this process: in the first experiments, only 0.05% of the cells were reprogrammed.

Then these numbers grew, in some cases even up to 10%, but most cells still did not respond to the technique (see S. Yamanaka, 2009. Elite and stochastic models for induced pluripotent stem cell generation). Yamanaka attempted to explain this through the landscape of Waddington (Fig. 3). Let's assume that we are trying to throw a ball back up the mountain. There may be several obstacles on its way: it may not reach the top and slide down completely (in this case, this means apoptosis, that is, intentional death of the cell, arrow 4), it may roll to the top and not stay there (then the cell will differentiate again in a random direction, arrow 2) or it may fly in the wrong direction and jump into the next track (differentiate into another type, arrow 3). And only in those rare cases when the ball reaches the top and is held there, the cell is reprogrammed and becomes stem (arrow 1).

Fig. 3. Application of the Waddington landscape to cell reprogramming. 1 – full return to the embryonic state. 2 – failed reprogramming. 3 – transition to another cell type. 4 – cell death. Image from the article: S. Yamanaka, 2009. Elite and stochastic models for induced pluripotent stem cell generation.

But the adventures of the ball on the slide do not end there. Until recently, only two main methods of working with stem cells were known: differentiation and reprogramming. Differentiation protocols consisted of sequential action on cells by substances "directing" them in one direction or another. The set of these substances was determined experimentally, based on real development processes. If retinoic acid is needed to differentiate the cells of the nervous system at some stage in the mouse embryo, then in the laboratory it can be added to the culture medium to produce neurons. Now imagine that we want to get, for example, a culture of human neurons. You can take his skin cells, reprogram them to embryonic stem cells, and then differentiate them into neurons. It turns out to be long and inefficient, so the search for a shorter path has been going on for a long time.

There were two short cuts – direct differentiation and transdifferentiation. In the course of direct differentiation, it is proposed to obtain neurons from embryonic stem cells without intermediate stages. Transdifferentiation involves obtaining neurons directly from skin cells. Note right away that cell types are given here as examples, in reality they can be almost anything. These short paths are carried out in one way: in the culture of the original cells (stem or differentiated), the expression of transcription factors characteristic of the desired cell type (in this case, neurons) is triggered. The effectiveness of these methods is still low, but both of them work. Therefore, many questions arise: how functional are the cells obtained in this way? What mechanisms underlie accelerated differentiation? What happens to the intermediate stages, do they disappear completely or are they partially preserved? These questions were asked by the authors of the article under discussion (Fig. 4).

Fig. 4. Research questions of the authors of the article under discussion. a) How do intermediate stages work in the case of direct differentiation? Possible options: all stages are present, some stages are skipped, alternative stages appear. b) Which of the intermediate stages disappear? Possible options: the earliest, the latest, the whole way is different from the classic. c) Are full-fledged functional cells obtained as a result of direct differentiation? Possible options: full or partial achievement of the effect. An image from the article under discussion in eLife.

The authors worked with mouse embryonic stem cells. They launched in parallel a standard protocol of sequential differentiation and direct differentiation into motor neurons of the spinal cord. To track the stages of the cells, they analyzed the RNA in individual cells at the early (4-5 days) and late (11-12 days) stages of differentiation. Then, the cells from which too little RNA was obtained were excluded, and those in which increased expression of mitochondrial genes associated with stress was detected: these cells may have developed abnormally. In the remaining cells, genes were calculated, the expression of which changed statistically significantly during differentiation.

This allowed us to determine the main stages through which the cells passed. With the standard differentiation protocol, the path turned out to be as follows: embryonic stem cells – common neural precursors – cells of the posterior nervous system – cells of the abdominal side of the posterior – precursors of motor neurons – early motor neurons – late (mature) motor neurons. At the same time, several stages could be detected simultaneously in the culture, since cells differentiate asynchronously. The scientists found that when using the standard protocol, fewer cells reached the late stages than with direct differentiation, and also more cells deviated from the intended path and turned into other types. At the same time, with direct differentiation, two stages could not be detected: the cells did not gradually acquire the properties of the posterior and abdominal parts of the nervous system (Fig. 5). At the same time, the initial and final stages in both protocols turned out to be very similar. The result was complete and functional motor neurons.

Fig. 5. Sequential stages that cells undergo during direct (b) and gradual (c) differentiation. Statistical indicators reflecting the expression of groups of genes characteristic of different stages are deposited along the axes. Cells with similar values of these indicators were considered to be at the same stage. Stage designations: ESC – embryonic stem cells, NP – common neural precursors, PNP – cells of the posterior nervous system, VNP – cells of the abdominal side of the posterior, MNP – precursors of motor neurons, EMN – early motor neurons, LMN – late (mature) motor neurons. An image from the article being discussed in eLife.

This difference of direct differentiation from the standard one clearly shows us the difference between the development of a cell in an embryo and in culture. In an integral system, such as an embryo, the layout of the building plan occurs at the early stages, and only after that the final differentiation begins. This is probably necessary for adequate interaction between tissues and different types of cells in the embryo. At the same time, these stages are not critical for the subsequent formation of a functional neuron, if we are talking about a culture where cells do not interact with the environment. During direct differentiation, the cells went through a stage uncharacteristic of the standard protocol: forebrain genes were included in them. However, by the later stages their expression disappeared, and the final stages of differentiation in both protocols turned out to be very similar.

The technology of direct differentiation raises many questions and disputes. Despite its effectiveness, before using it in practice, it is necessary to make sure that "non-standard" ways of cell development do not introduce any side properties into them. The authors of the article under discussion believe that the stage of a terminally differentiated cell belongs to the so-called "pools of attraction" (see Basin of attraction), that is, stable states to which all slightly different states strive and which are resistant to moderate environmental fluctuations. This means that if, as a result of differentiation, the expression of key transcription factors is achieved in any way, then the cell will somehow come to its final state. And the probability that she will take a different path at the last moment is extremely small. If this is true, then we can expect that short differentiation paths will soon replace long and traditional ones, increasing the efficiency of the process and bringing us closer to using these technologies in medicine.

Source: Briggs et al., Mouse embryonic stem cells can differentiate via multiple paths to the same state // eLife. 2017.

Portal "Eternal youth" http://vechnayamolodost.ru