Human development

Advances in mini-brain technology

Two studies integrate cutting-edge techniques to grow and analyse 3D cultured tissues that resemble human brain structures, enabling examination of how brain regions interact and neurons mature. See Articles p.48 & p.54

Over the past few years, the production of human brain-like tissue from stem cells in 3D cultures1,2 has allowed sophisticated analyses of how the brain develops, how its development has changed during evolution and how it is affected by disease3,4,5. But it has remained unclear precisely what cell types arise in these brain 'organoids', how much individual organoids vary, and whether mature neuronal networks can form and function in organoids. Using a combination of sophisticated techniques, two papers6,7 in this issue describe key steps towards addressing these questions.

There are two general strategies for growing human brain organoids. In the first strategy, pluripotent stem cells (PSCs, which can give rise to any cell type in the body) are guided to form a layer of stem cells called a neuroepithelium that can make neurons. The neuroepithelium is left to develop alone, and this can result in the generation of multiple brain regions (Fig. 1a). This self-patterning strategy offers the potential to understand how brain regions self-organize and interact. However, there are often substantial differences between individual organoids, and between batches grown separately.

Figure 1: Two routes to a mini-brain.

Structures called organoids that resemble regions of the embryonic human brain can be grown in vitro from human pluripotent stem cells (PSCs), which can give rise to every cell type in the body. a, Quadrato et al.6 induced PSCs to form a cell layer that self-patterned into organoids composed of multiple brain regions. Measurements of neuronal activity revealed that the organoids contain cells that respond to light in a similar way to photoreceptor cells of the retina, demonstrating that neurons can mature in organoids. The authors isolated single cells and analysed the gene-expression profiles (transcriptomes) to determine the range of cell types in the organoids. b, Birey et al.7 used signalling molecules to direct the development of 3D structures called spheroids that resemble two regions of the forebrain (the ventral forebrain and dorsal pallium). Spheroid fusion led to the formation of forebrain-like organoids. Imaging revealed that neurons called interneurons migrated from the ventral to the dorsal region, providing information on brain-region interactions. These authors also performed single-cell transcriptomics.

The alternative strategy is to use signalling molecules to control patterning of the neuroepithelium so that a defined region forms — the forebrain or hypothalamus, for instance (Fig. 1b). This technique might increase reproducibility, but researchers have yet to define all of the signals that create each subregion of the brain. Moreover, it is unclear whether certain regions can form in the absence of adjacent structures.

Quadrato et al.6 (page 48) set out to examine the composition and functionality of maturing brain organoids in detail. The authors modified a self-patterning protocol3 to reduce cell death, which can occur towards the centre of organoids owing to a lack of oxygen. Organoids grown using this modified method progressively matured for more than nine months, making it possible to study how neurons develop over a time equivalent to that of human gestation.

The researchers analysed the gene-expression profiles (the transcriptomes) of more than 80,000 cells from 3- or 6-month-old organoids — the most comprehensive single-cell analysis of organoid composition performed so far. These data revealed diverse cell populations from different brain regions. Focusing on retinal cell types, the investigators demonstrated that their organoids contained almost every cell type found in this tissue in vivo. They also found evidence that neurons mature over time, beginning to express genes that mediate synaptic connections with other neurons.

How mature do these neurons become? Electron microscopy on serial slices of tissue revealed that an 8-month-old organoid contained a density of synapses approximately equivalent to that in the cortex of an embryonic human brain. Individual neuronal projections (dendrites) often made multiple synapses, suggesting that complex networks form by this stage. Quadrato et al. found that these synapses could fire, and periods of coordinated firing indicated that organoids contained active neuronal networks.

Finally, the authors showed that a subpopulation of neurons had lowered firing rates after exposure to light. Coupled with transcriptomic data suggesting that organoids contained light-sensing retinal neurons called photoreceptor cells, these data indicate that organoids could be used to investigate how neuronal networks are modulated by physiological stimuli.

It is of note that Quadrato and colleagues found extensive differences in cell composition between organoid batches. This highlights the need to control organoid engineering so as to improve reproducibility when studying disease mechanisms, for example. In the second study, Birey et al.7 (page 54) take a leap in this direction.

The authors used a controlled-patterning protocol to generate 3D structures, which they called spheroids, that resemble one of two forebrain regions: the ventral forebrain, which produces inhibitory neurons called interneurons; and the dorsal pallium, which contains excitatory neurons. Single-cell transcriptomics showed that the spheroid cells were remarkably similar to those from corresponding regions of the human fetal brain.

Next, Birey and colleagues placed spheroids of different types next to each other, and allowed them to fuse over a few days to form forebrain-like organoids. Interneurons from the ventral spheroids migrated into the dorsal regions — similar to the migration route taken by these interneurons in vivo. After migration, the interneurons matured. The authors detected stronger neuronal firing in fused organoids after interneuron migration than before.

The researchers used live-cell imaging to track individual interneurons, comparing their behaviour with that of equivalent cells in the human and mouse brain. Certain human-specific aspects of migration were accurately recapitulated in the fused organoids, presenting exciting possibilities for studying the genes that mediate migration. Indeed, Birey et al. next used their system to study the neurodevelopmental disorder Timothy syndrome, which is associated with mutations in a protein that regulates interneuron migration8.

The authors produced fused forebrain organoids from PSCs that had been generated by inducing mature differentiated cells from patients with Timothy syndrome to become pluripotent. The patient-derived interneurons migrated less efficiently than controls. This is probably due to a defect in the interneurons themselves, rather than in the dorsal environment, because migration remained defective when the researchers fused patient-derived ventral spheroids to wild-type dorsal spheroids. Birey et al. have therefore succeeded in generating a controlled 3D system that has both excitatory and inhibitory neuronal activity, and that can effectively model an interneuron migratory disorder.

Both studies reveal the power of high-throughput, single-cell transcriptomics to quantify cell-type diversity in brain organoids. These methods will be valuable for studying how cell composition and gene-expression networks are dysregulated during disease. Nevertheless, it is unclear how transcriptomes will vary between organoid batches, between different stem-cell lines generated from the same patient, and between patients with the same disease. This is particularly important when comparing patient and control organoids, and should be addressed in detail.

It was known4 that organoid neurons could create synapses. However, serial electron microscopy is a new and elegant approach to studying detailed synaptic structures in organoids, and, it is hoped, to studying the complete map of neuronal networks. It remains to be seen how neurons 'decide' to connect with one another in these complex and heterogeneous tissues, and whether the connections mirror those in the developing human brain.

Much work shows that brain tissue engineered from induced PSCs can accurately model neuronal-progenitor behaviour9. The new studies showcase how brain-organoid technology can be improved to study the interactions between brain regions and the way in which mature neurons might function in a coordinated network. Time will tell just how much information these mini-brains can provide. Footnote 1


  1. 1.

    See all news & views


  1. 1

    Sasai, Y. Nature 493, 318–326 (2013).

    ADS  CAS  Article  Google Scholar 

  2. 2

    Lancaster, M. A. & Knoblich, J. A. Science 345, 1247125 (2014).

    Article  Google Scholar 

  3. 3

    Lancaster, M. A. et al. Nature 501, 373–379 (2013).

    ADS  CAS  Article  Google Scholar 

  4. 4

    Qian, X. et al. Cell 165, 1238–1254 (2016).

    CAS  Article  Google Scholar 

  5. 5

    Mora-Bermúdez, F. et al. eLife 5, e18683 (2016).

    Article  Google Scholar 

  6. 6

    Quadrato, G. et al. Nature 545, 48–53 (2017).

    ADS  CAS  Article  Google Scholar 

  7. 7

    Birey, F. et al. Nature 545, 54–59 (2017).

    ADS  CAS  Article  Google Scholar 

  8. 8

    Bortone, D. & Polleux, F. Neuron 62, 53–71 (2009).

    CAS  Article  Google Scholar 

  9. 9

    Kelava, I. & Lancaster, M. A. Cell Stem Cell 18, 736–748 (2016).

    CAS  Article  Google Scholar 

Download references

Author information



Corresponding authors

Correspondence to J. Gray Camp or Barbara Treutlein.

Related links

Related links

Related links in Nature Research

Developmental neuroscience: Miniature human brains

Systems neuroscience: A modern map of the human cerebral cortex

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Camp, J., Treutlein, B. Advances in mini-brain technology. Nature 545, 39–40 (2017).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.


Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing