Researchers studying neural development are plagued with an ever-present difficulty: how does one study the developing human brain without using invasive methods or being restricted to post-mortem tissue? Due to the complexity of human brain development, research has been able to answer few questions about neural developmental defects beyond those observable in animal models or in human cell cultures. This task is further complicated by the inconvenient reality that brain development in humans differs in key ways from brain development in mice, and that small, two-dimensional cultures have rarely recapitulated the organization of neural tissue in normal development.
Fortunately, a new and innovative technique published in Nature has recently attracted much attention for enabling a more precise model of human neural development. Published in the September 19th issue, this report from the lab of Austrian researcher Juergen Knoblich details a three-dimensional culture system using stem cells to generate miniature brain-like tissues, called ‘cerebral organoids’ .
Putting Stem Cells Back in Control
Stem cells, or cells of the body that give rise to many, more specialized cells, can be used to model human development and are anticipated to enable therapies for a host of developmental and degenerative conditions. Embryonic stem cells (ESCs) and induced pluripotent stem cells (iPSCs) — ESC mimics produced from adult tissues — are hailed for their pluripotency, or ability to differentiate into each cell type in the body. Yet, for all the purported promise of stem cells, they are notoriously difficult to manipulate and guide into specialized cells that yield organized tissues.
Exploiting the incredible self-organizing capacity of ESCs as demonstrated by development in the womb, Lancaster and colleagues eschewed common practice by supplementing the cells with only minimal amounts of growth factors, the chemicals that are typically added to stem cell media to forcibly direct the generation of specific cell types. Instead, the researchers established suitable growth conditions for the cells to influence their own development through their own intrinsic and secreted cues, enabling the production of tissue assemblages up to 4 mm in diameter that exhibited readily recognizable brain regions of the cerebrum (the largest and most developed part of the brain), complete with regional tissue specification and distinct brain-cell subtypes within each region (Figure 1A). In short, the lab has generated the closest thing yet to a mini-brain in a dish, approximating the organization and differentiation of neural stem cells into mature brain cells during embryonic development. Because the cerebral organoids lack blood vessels and the same intricate connectivity of fully developed brain tissue, they can only be used to model early embryonic development, but also sidestep the ethical questions that would accompany creation of a more complex and potentially conscious nervous system.
Figure 1. Cerebral organoids introduce new methods of modeling defects in neural development, particularly those of the cerebral cortex. (A) Organoids are composed of neural progenitors (red) and neurons (green), as well as other cell types (blue marks all cells). In addition, organoids develop tissues such as neocortex, choroid plexus, and ventricle. (B) The cerebral organoid culture system begins with dissociated pluripotent stem cells which are first directed them to a neural fate. Neuroectoderm (early neural stem cells) are then embedded in Matrigel, after which they are cultured in a spinning bioreactor and direct their own development in the absence of most exogenous growth factors. Modified from Lancaster et al. (2013).
The Complexity of Brain Development
To fully recognize the significance of such a discovery, it is important to understand the basics of how the brain develops (Figure 2). In its most primitive form just prior to the fourth gestational week, the brain exists as a simple tube composed of a single layer of neural stem cells, called neuroepithelial cells, that enclose a fluid compartment called the ventricular system . To generate the many cell subtypes and diverse regions of the cerebrum, neuroepithelial cells proliferate to expand the neural stem cell pool, and then transition into radial glia progenitor cells (RGCs), which themselves divide either to produce more progenitors or to generate newly-born brain cells, called neurons. As they are born, neurons migrate to their proper location in the brain, often guided by signals from other cells. In the neocortex (a recently-evolved, layered region of the cerebrum and the seat of many higher functions in mammals), neurons migrate past one another up fibers that stem from RGCs, populating distinctive layers of the cortical plate .
Figure 2. A basic understanding of the development of the cerebral cortex is necessary to understand why cerebral organoids represent such as step forward in modeling cortical development. (A) The early neural tube is composed of a single layer of pseudostratified (having the appearance of a layered tissue) neuroepithelial cells. Cells migrate up and down through the tissue, dividing at the ventricular border. (B) Neuroepithelial cells in the neural tube divide several times before transitioning to radial glia progenitor cells, which proliferate further (more in humans than in mice) before generating newborn neurons. These neurons migrate up the radial glia processes and take positions in distinct layers of the cortex (forming layers 6 to layer 2, inside out). This same inside-out organization is recapitulated in cerebral organoids.
The Knoblich study is not the first instance of generation of organized brain regions outside of the human body, but it is certainly the most complex. In 2012, Japanese researcher Yoshiki Sasai demonstrated the self-organization and morphogenesis of the human retina (a little-recognized brain region in the eye where incoming light is encoded into electrical signals) from human ESCs in a three-dimensional culture . However, even though this and other discrete brain regions have been created in a dish, never before have discrete yet interdependent brain regions been created together in one culture system.
Using 3-D Culture to Make Cerebral Organoids
To generate the cerebral organoids, Lancaster and colleagues guided ESCs into a neural stem cell fate and embedded the fledgling tissue into droplets of a structural protein mixture called Matrigel (Figure 1B). The Matrigel droplets were then added to culture media in a spinning bioreactor that applied rotational motion to assist in nutrient absorption. Here, the organoids developed defined brain regions within 20-30 days and reached maximal size by 2 months. Extraordinarily, most organoids established a fluid-filled cavity resembling the ventricular system, complete with radial glia progenitors in their correct location adjacent to the ventricular space. Furthermore, organoids developed a neocortex with proper polarity and specification of layers, as well as a hippocampus (critical for storing memories), choroid plexus (which secrete fluid into the ventricles), and immature retina.
Using Cerebral Organoids to Model Disease
The cerebral organoid culture method introduces a promising system for modeling previously intractable developmental disorders. Using this model, patient-derived iPSCs — generally skin cells that have been artificially reprogrammed to restore pluripotent stem cell identity — can be used to recapitulate complex developmental defects in a dish. To illustrate this concept, the authors fashioned iPSCs from a patient with severe microcephaly, a developmental defect characterized by greatly reduced head and brain size, which, for this patient, was conjectured to be caused by a genetic mutation in the CDK5RAP2 gene. Cerebral organoids generated from these iPSCs exhibited atypically few radial glia progenitor cells, leading to a deficiency of mature neurons reminiscent of the patient’s microcephaly. Furthermore, reducing or replenishing CDK5RAP2 protein in patient-derived iPSC organoids was able to diminish or enhance, respectively, the level of neural progenitor proliferation. This set of experiments with iPSC-derived cerebral organoids lends novel insight into the mechanisms behind CDKRAP2-associated defects, and supports the idea that insufficient production of neural progenitor cells could underlie microcephaly.
The CDK5RAP2 mutation has been studied before in animal models . However, the mutation affects one aspect of human brain development in particular that doesn’t occur to the same degree in mice: the extensive expansion of the neural progenitor pool prior to neurogenesis, which in turn enables more neurons to be created. When Knoblich’s group generated mouse cerebral organoids, they found that far fewer outer radial glia cells (oRGs), a neural progenitor subtype, were generated in mouse organoids than in the human organoids. In addition, all mouse neural progenitors generally underwent fewer divisions to expand the stem cell pool prior to neuronal differentiation. The CDKRAP2 mutation causes premature differentiation of these neural progenitors into neurons, and thus has a greater effect on humans than on mice.
For this reason, the organoid approach is a game-changer for studies of human neural development. Previously, studies of human-specific defects could not be recapitulated using controlled, empirical methods. To circumvent this difficulty, animal models such as mice, zebrafish, and fruit flies have been frequently utilized to study complex developmental defects. Animal models can be perfectly satisfactory to recapitulate some human diseases, but for developmental defects that are unique to humans, it becomes necessary to use human cells. The cerebral organoid approach provides a fantastic resource to do just that.
Mark Springel is a research assistant in the Department of Pathology at Boston Children’s Hospital.
 Lancaster MA, Renner M, Martin C, Wenzel D et al. (2013). Cerebral organoids model human brain development and microcephaly. Nature 501: 373–379
 Gotz M and Huttner W (2005). The cell biology of neurogenesis. Nat Rev Mol Cell Biol 6(10): 777-88
 Nakano T, Ando S, Takata N, Kawada M et al. (2012). Self-formation of optic cups and storable stratified neural retina from human ESCs. Cell Stem Cell 10(6): 771-85
 Barrera JA, Kao LR, Hammer RE, Seemann J, et al. (2010). CDK5RAP2 regulates centriole engagement and cohesion in mice. Dev Cel 18(6): 913-26
 Lab-Grown Model Brains (The Scientist Mag):
 Nature News Feature (Nature): http://www.nature.com/news/tissue-engineering-the-brainmaker-1.11232