Are there any neural stem cells in the brains | of adults? Image source: pixabay.com
Introduction
Are there still neural stem cells in the adult brain, and can these neural stem cells continue to generate new neurons? For more than 10 years, people in the industry generally believe that there exists; however, some studies in recent years have believed that there is no existence.
Generally speaking, proving the existence is relatively simple; proving that it does not exist may have a larger workload, requiring more stringent identification standards and withstanding the test of statistics.
In any case, when the neurodevelopment is over, the mature human brain has a total of about 80 billion neurons and 80 billion glial cells, and finding neural stem cells in all of these cells is challenging.
Related Reading
Scramble | Fudan neuroscientists: There are no more newborn neurons in the adult brain
Written by| Qian Cheng, Zhou Fengquan (Johns Hopkins University)
Responsible editor| Di Li Hui
● ● ●
1
To heal and get smarter
Whether there are neural stem cells in the adult brain, scientists are concerned about this question, there are two simple reasons.
First, if neural stem cells are still present in the adult brain, and these neural stem cells can be stimulated by some means to divide and proliferate, and re-differentiate into neurons, it may be a very effective new treatment for nerve damage and neurodegenerative diseases.
Neurons die after being damaged or invaded by disease. Many neurodegenerative diseases, such as Alzheimer's disease, Parkinson's syndrome and other diseases, the main pathological consequence is the death of neurons. The death of neurons is the leading cause of cognitive decline.
Because neurons cannot re-divide to produce new neurons, there is currently a lack of effective treatments for human brain damage or neurodegenerative diseases.
In addition, over the past 20 years or so, many laboratories have studied the function of neural stem cells in the brains of adult mice.
They found that there were a lot of neural stem cells in the hippocampus region of the brains of adult mice. These adult neural stem cells are similar to neural stem cells during neurodevelopment, can continue to divide, produce more neural stem cells, and can also differentiate into neurons and glial cells, and differentiated neurons can be integrated into the neural circuits that have been formed to function.
In addition, a series of other studies have shown that if mice are trained on exercise or intelligence, many of the mice's advanced cognitive functions, such as learning and memory ability, will increase significantly.
It was agreed that the hippocampal region of the mouse brain is a key brain region responsible for learning and memory, and further studies have shown that neural stem cells in the hippocampal region play a very important role in improving cognitive function in mice.
Based on these findings, it is believed that the neural stem cells present in the brains of adult animals and their ability to produce newborn neurons are also an important component of brain neuroplasticity.
It follows that if similar neural stem cells exist in the adult brain, they can also increase their cognitive abilities or become smarter through exercise or intellectual training.
2
How to identify human brain neural stem cells?
Samples of the human brain come from two main sources, the most common being brain tissue donated by the families of the deceased. In addition, some patients with brain diseases need to undergo partial brain tissue resection surgery during treatment, and the brain tissue that is removed is also another major source of samples with the consent of patients and their families.
Due to the peculiarities of the sample source, the possibility of any pretreatment of the sample is low, although it is low. To sum up, there are probably the following methods to find and identify human brain neural stem cells.
One is to directly search for neural stem cells.
Adult neural stem cells are basically similar to neural stem cells during embryonic development (including specific proteins, brain regions, and transcriptome characteristics). Neural stem cells can divide and contain some proteins (marker proteins) that are relatively specifically expressed in neural stem cells.
Second, it is possible to look for new neurons produced by the differentiation of neural stem cells. These newly generated neurons also express some relatively specific marker proteins compared to the original old neurons.
The third is to first label all cells that are dividing (such as Brdu, Bromodeoxyuridine, Chinese name is bromodeoxyadenadentine. It is an analogue of deoxythymidine in DNA that can be incorporated into new and formed DNA as cells divide, often used to detect dividing living cells), and then look for cells containing these markers and also expressing neuron-specific proteins over time. These cells are neurons that are produced by the differentiation of neural stem cells.
The above three methods are mainly done by antibody staining experiments.
Finally, by single-cell sequencing cells in specific areas of the brain, a transcriptome database of each cell sample can be obtained. By bioinformatics analysis of this database, it can help us identify cells with neural stem cell transcriptomes.
Although experimental methods have gone through several technical generations —from tissue slice in situ immunostaining, dispersed cell staining and flow cytometry counting, to bulk tissue RNA-seq of whole batches of cells mixed together, to single-cell (nuclear) RNA sequencing (see table below) — there is no consensus on the existence or absence of neural stem cells in the hippocampus region of the adult brain, and whether endogenous adult neurogenesis exists.
Click to view a larger image
Here we will summarize the conclusions of the listed articles as "having" or "not" finding adult neonatal neurons. As for the article that concludes that there is "yes", the information about whether the number of existences is more or less, and the background of the disease of adult human samples, you can go to the original text to find. Due to time constraints, we did not compare the technical details of sample preparation in these staining experiments one by one, and whether antibody selection and optimization were reasonable. For example, we do not discuss in detail whether neural stem cells or newborn neurons in the sample are exposed to specific antigens, or whether antibodies bind to other proteins non-specifically. Similarly, the impact of BrdU on 14-carbon isotope analysis is not discussed here, whether BrdU can be incorporated in small amounts of non-dividing cell DNA, or the technical details of biological sample preparation.
There are always flaws in science and technology, but as more data continues to emerge and a number of different technologies complement each other (such as FISH RNAscope (RNA in situ hybridization) and multiplex immunostaining (multiple antibodies combined), the technical barriers will become smaller and smaller.
If the staining strategy is adopted, then each cell is infected and the cell is captured, and the rest is to assign each cell identity to neural stem cells by what criteria.
If you are wrong, you will create two results, either false positives or false negatives. If you don't consider the technical barrier that the poor quality of the antibody binds to other proteins other than the target target, the remaining problem is —
◢ How many neural stem cells/neonatal neuron-specific identity markers are needed?
◢ Two or three is enough?
◢ Is the specificity of the selected marker strong enough?
◢ Are neural stem cells/neonatal neuron marker proteins also expressed in some glial cells?
◢ If expressed, can glial cell markers be excluded by looking at them?
◢ Can the prior knowledge obtained from the detection of mice, monkeys, and juvenile humans be completely copied and used in the exploration of adult brain neural stem cells?
3
A better way
In addition to the classic cell-marker protein staining technique, we believe that studying cell identity through new technology single-cell sequencing should be a more accurate method.
So, can single-cell (nuclear) sequencing capture those adult brain neural stem cells (if any) found in staining experiments above (see table above)?
Theoretically, it should be fine.
Staining experiments are used to detect proteins directly by antibodies, while nucleic acid sequencing is the messenger RNA that detects the coding proteins. The conventional sequencing depth set by default usually measures an average of 3,000 or 4,000 expressed genes per cell (nucleus), although the added overhead can be measured even deeper. Staining experiments typically stain up to 3 or 4 proteins, and usually these protein markers are not expressed too low, so the RNA encoding them should be able to be sequenced by single cells (nuclei).
If only 3 or 4 gene expressions are concerned, then the sequencing of cells (nuclei) is, in a way, equivalent to identifying neural stem cells for 200,000 cell samples by detecting 3 or 4 cell markers (in addition, in addition, in addition to identifying cell types, single-cell sequencing results can also be inferred from big data analysis, and the ability to divide is also one of the characteristics of stem cells).
So how exactly do you identify a cell type by single-cell (nuclear) mRNA sequencing?
First, using the transcriptome of 3,000 or 4,000 expressive genes measured on average per cell, we can group cells that are similar to each other into one category through big data analysis in a relatively high number of dimensions, that is, cluster analysis.
With a sufficient number of captured cells (nuclei) and a good sequencing depth, each type of cell can be subdivided into subtypes. Although the software or computational methods used may vary, clustering is itself an unbiased process based on correlation.
But for each cluster of cells that are separated, the identification of cells is ultimately human, and it is a step that relies heavily on prior knowledge.
Specifically, through data analysis, it can be calculated that each cluster expresses 50 to 100 genes with significantly higher expression, and then determine the identity of the cluster according to several markers of known cell types in these genes, and finally can be combined with FISH or antibody staining to confirm.
Of course, there is hope for the discovery of novel, more pure cellular identity markers in these 50 to 100 genes. Similarly, clustering after single-cell sequencing can re-verify that certain traditional cellular identity markers are indeed not expressed in other cell types (cell-specific). For example, the 2021 Neuron journal article [1] found that in the adult brain hippocampus, the commonly used newborn neuron marker protein Dcx is not just a marker of immature newborn neurons.
Since human brain samples are not easily sourced, in most cases samples can only be taken at a single point in time, rather than in animal models where samples from several points in time are available. But even at a single point in time, the state of each cell within the same tissue varies slightly. In other words, in a group of cells there will be subpopulations of cells at different stages of a biological process.
Existing big data analysis techniques can find these cell subsets from single-cell (nuclear) sequencing data sampled at a single time point through an analytical method called Pseudo-time trajectory, thus mimicking the entire biological process.
For example, the 2021 Neuron journal article [1] found that by analyzing the single-cell sequencing results of adult mice, pigs, and rhesus macaques, it is possible to depict a process of neural cell differentiation and growth from the neural intermediate progenitor cells, to the newborn neurons, and then to the mature neurons (granule cells). Under the same conditions, however, the researchers could not detect this cellular process in the hippocampus region of the adult human brain. Based on this result, the study concluded that neural stem cells were not present in these human brain samples.
The two single-cell nuclear sequencing articles [1,2] published in the journal Neuron in 2021 both focused on exploring neural stem cells and newborn neurons in the hippocampus region of adults. In short, none were found.
The exploration of neural stem cells and neurogenesis processes in the dentate gyrus region of the adult hippocampus is due to the fact that earlier experiments have found significant traces of newborn neurons, or at least neuronal growth processes, in adult mammals, such as mice in rodents, non-human primates, and juvenile human hippocampal dentate gyrus.
As for neural "stem cells" in the adult brain, if they exist, it is generally believed that the regions and other properties in which they exist are similar to those found in other animal models and in the brains of young humans. Whether this preconceived notion is correct or limits our horizons, because there is no prior basis for answering, but it is worth paying attention to.
4
The challenge of single-cell sequencing
Single-cell (nuclear) mRNA sequencing is also not flawless, and often the disadvantages come from the advantages.
Single-cell sequencing also relies on prior knowledge to assign qualitative values to cell identity. After selecting several parameters during analysis, clustering can be automatically calculated by the computer. But identifying each cluster as a cell type, subtype, or state still requires artificial assignment according to a priori markers.
For example, if you blindfold to touch such as 200,000 animals, each with 3,400 or 4,000 characteristic values (or even more to reach tens of thousands of characteristics), you can divide the elephants into a pile, giraffes, sheep, cats and dogs, poultry, fish in piles, the analysis of the finer can also continue to separate cats and dogs, chickens, ducks and geese.
We are able to succeed in assigning identities because of prior knowledge, and we already know these animals. Imagine if we didn't know the alpaca before, and the number was not large, then it is very likely that 7 or 8 alpacas will be divided into a pile of sheep, that is, false negative: there is no alpaca.
Similarly, what if adult brain neural stem cells are not similar to these prior knowledge of mice, pigs, monkeys, and human infancy?
For example, the 2021 Neuron Journal article [1] These 200,000 nuclei have successfully captured neural stem cells, but they are not recognized (false negative), but they are hidden in the astrocytes, microglia, oligodens cells, interneuron groups.
Fortunately, once adult brain neural stem cells have been successfully captured by single-cell sequencing, with technological advances that can measure higher throughput cells and the depth of sequencing increases (each platform has such a new solution introduced), this stem cell cluster should theoretically emerge. While perhaps the cluster will share many marker genes with, for example, glial cells or endothelial cells, the sequencing resolution should be sufficient to find a difference.
More importantly, as a process, the trajectory of adult neurons should be present, measurable, and clear.
Because if it is eventually found that some astrocytes or microloglia cells have neural stem cell function, they cannot be instantaneously converted to the exact same as some original adult brain cells, so it should be possible to identify them through transcriptome analysis.
Even if clustering and trajectory analysis of transcriptomes and proteomes cannot assign values to identify adult neural stem cells (if present, and captured), clustering and trajectory analysis based on chromatin structure, including enhancer enhancer characteristics, and so on. Under the trend of multi-omics, single-cell (nuclear) DNA methylation characteristics, lncRNA characteristics, and single-cell ChIP-seq will also enter the market sooner or later to become useful tools.
If the above methods still can't find adult neural stem cells, what does it mean?
One possibility is that stem cell-like and neuron-like germination processes in the adult brain are resting or inhibited in a normal physiological state.
For example, a recent study found that neural stem cells senesculating in the hippocampus region of adult mouse brains inhibited the division and differentiation functions of normal stem cells near them. Removing aging stem cells by chemical means can greatly promote the division and differentiation of normal neural stem cells.
Similarly, neural stem cells in the adult brain may also need to be activated by certain conditions, such as injury, a disease, a specific genetic, or environmental background. This kind of regulation is not necessarily linear or perhaps a dynamic system, so the future is still bright.
As more data are produced, especially those that stain to find adult brain neural stem cells, multi-omics single-cell (nuclear) sequencing can be applied to confirm their findings.
The old bulk tissue mRNA-seq is also a treasure trove (usually in geo databases), because many samples with some kind of background or damage may not be able to find the same one for single-cell sequencing. Combined with the single-cell sequencing marker library, there are now many established data analysis processes [3] that can deconvolution of burk-seq to see the different genes of each cell component.
In addition, other defects in single-cell sequencing, such as the histizymatic digestion process, can alter cell state and transcriptome, or cause transcripts to be lost. Our own experience has also found that the enzymatic digestion process of brain tissue in adult animals, if not optimized, can cause a large number of cells to be lost.
Some of the new sequencing schemes may be able to overcome these drawbacks and make the measurement results more authentic.
For example, spatial sequencing can convert "pixels" of tissue slices directly in situ into "single cells" for resequencing without enzymatic digestion. Another example is the single-cell sequencing by tissue section FISH RNAscope verification to find transcription products that are not affected by sample digestion.
5
The jury is still out
There are about 80 billion neurons in the adult brain, as well as about the same or slightly more number of glial cells and other cells.
Strictly speaking, to know whether this 160 billion cells per person has "have" or "does not" neural stem cells and adult newborn neurons, if you find a conclusion to "yes", then a key question is how to determine whether the cells found are neural stem cells? Approximately how many neural stem cells make up this 160 billion cell population?
If not found, conclude "no", then an important question is whether the coverage and quantity of samples are sufficient? For example, with single-cell sequencing, 200,000 cells are sampled at a time, how many times do you have to measure to draw conclusions? Consider whether it can withstand the test of statistics.
In the era of multi-omics big data, let the bullets of adult brain neural stem cells fly a little longer.
Background tips
The formation of all animal brain tissue must go through a neurodevelopmental process. The main cell type of neural tissue in the early stages of neurodevelopment, including the brain, is neural stem cells.
These neural stem cells can divide and proliferate, and some of them can also stop dividing and differentiate into two main types of cells in the nervous system: neurons and glia. Differentiated neurons grow axons and dendrites that are then connected to each other to form complex neural circuits. Neural circuits formed by neurons, by interacting with various types of glial cells, form a complete brain neural network system that controls various types of neural functions. These neurological functions include sensory functions such as hot and cold, itchy, smelling, taste, hearing, visual function, motor function, and various advanced cognitive functions. When neurodevelopment is over, the mature human brain has a total of about 80 billion neurons and 80 billion glial cells, meaning that finding neural stem cells in all of these cells is challenging. The differentiation of neural stem cells into neurons is an irreversible process, and neurons do not have the ability to re-divide to produce new neurons, so they are not replenished after death. Over a person's lifetime, the number of neurons gradually decreases with age. Unlike neurons, glial cells can return to a divisional state and produce new glial cells under certain conditions, such as after being damaged. In lower vertebrates, such as zebrafish, certain glial cells, such as Muller cells in the retina, return to the state of neural stem cells after injury, resulting in the creation of new neurons and glial cells. This function of glial cells is generally considered to be completely lost in mammals. After the end of brain tissue development, in order to ensure that its internal neural network system is in a relatively stable state, so that it can exercise its function normally, mature neurons generally lose the internal ability to grow nerve fibers, and the external environment in which neurons are located is also a relatively inhibited state. Internal and external factors work together to maintain the stable operation of neural networks, which is also the main reason why mammalian brains cannot regenerate and repair damaged neural circuits after being damaged or invaded by disease. Mature neural networks are not completely fixed, and all neurons have a certain degree of neural plasticity, that is, the connections between neurons can be adjusted to a certain extent (such as connection density and strength, etc.). Studies have shown that neuroplasticity has a great effect on regulating various types of nerve function or repairing nerve damage.
In general, neuroplasticity declines with age, so humans in childhood can have a better ability to learn and master new things, such as learning multiple languages. On the other hand, brain tissue is more likely to be repaired in childhood after it is damaged or invaded by disease.
bibliography:
1. Frantic D, Skarica M, Ma S, Arellano JI, Tebbenkamp ATN, Choi J, Xu C, Li Q, Morozov YM, Andrijevic D, Vrselja Z, Spajic A, Santpere G, Li M, Zhang S, Liu Y, Spurrier J, Zhang L, Gudelj I, Rapan L, Takahashi H, Huttner A, Fan R, Strittmatter SM, Sousa AMM, Rakic P, Sestan N. Transcriptomic taxonomy and neurogenic trajectories of adult human, macaque, and pig hippocampal and entorhinal cells. Neuron. 2021 Nov 18:S0896-6273(21)00866-7. doi: 10.1016/j.neuron.2021.10.036. Epub ahead of print. PMID: 34798047.
2. Ayhan F, Kulkarni A, Berto S, Sivaprakasam K, Douglas C, Lega BC, Konopka G. Resolving cellular and molecular diversity along the hippocampal anterior-to-posterior axis in humans. Neuron. 2021 Jul 7;109(13):2091-2105.e6. doi: 10.1016/j.neuron.2021.05.003. Epub 2021 May 28. PMID: 34051145; PMCID: PMC8273123.
3. Le T, Aronow RA, Kirshtein A, Shahriyari L. A review of digital cytometry methods: estimating the relative abundance of cell types in a bulk of cells. Brief Bioinform. 2021 Jul 20;22(4):bbaa219.
Plate editor| Lucas