The brain is the most complex organ in the body, controlling our highest functions, as well as regulating myriad processes which incorporate the entire physiological system. There is a significant risk that a novel therapeutic agent might impact brain structure and function, resulting in serious pathologies and even death. Therefore,
Early stages of drug discovery are primarily reliant upon the use of in vivo and in vitro animal models. There are considerable costs involved with animal work, as well as relevance and ethical concerns; motivating research into human or in silico alternatives. Whilst the development of existing medicines and their safety profile would not have been possible without animal studies, there are obvious shortcomings with these methods. In addition, extrapolation between species can be difficult due to morphological and metabolic differences. How then can we generate human drug screening models without using techniques which are invasive to humans? The answer could lie with induced pluripotent stem cells (iPSCs) .
iPSCs were discovered in 2006 by Takahashi and Yamanaka , who found that ectopic co-expression of four transcription factors governing pluripotency (c-Myc, Oct4, Klf-4 and Sox2) with mouse fibroblasts successfully reverted fibroblasts into a state of pluripotency ( 1 ) . In the following years, protocols were developed using human fibroblasts and for conversion of iPSCs into a variety of adult cell types. Work by Chambers et al. (2009) ( 2 ) , and later by Shi et al. (2012) ( 3 ) , provided robust and efficient neural induction protocols, enabling conversion of neural stem cells into cortical neuronal cell types suitable for a range of purposes. Models of the cortex are desirable, as the cortex is the executive, integrative centre of the mammalian brain and hence, is intimately associated with neurodegenerative and other CNS disease progressions and adverse drug reactions.
Do they talk the talk, and do they walk the walk?
iPSC-derived models should be able to replicate the in vivo morphology and functionality of components as closely as possible. A cortical iPSC model should contain cortical markers identifiable by appropriate immunocytochemical methods. The cortex is a delicate six-layer arrangement of neurons and associated glial cells. However, many models omit glial cells from culture and the importance of co-cultures of both neurons and glia are often overlooked. Of course, neurons are the principal cell in the brain, despite being far less abundant than their glial counterparts. Neurons fire action potentials, facilitate synaptic transmission and respond to a variety of stimuli; whilst glia carry out their own transmission, communicate directly with neurons and importantly, affect drug metabolism. As such, generation of a cortical co-culture from neural stem cells, containing astrocytes as well as neurons, is a far more representative model of the human system .
Neuronal electrical activity is matured and enhanced by the presence of astrocytes (Odawara et al., 2014) ( 4 ) . Characterisation of electrical activity of cultures can be achieved with electrophysiological techniques. Traditional invasive techniques such as patch clamping are a very accurate measure of electrical activity and function, but are not suitable for high-throughput testing. Multi-electrode arrays (MEA) are a popular, effective, high-throughput system for assessing spontaneous and synchronous activity and drug responses of neural cultures . MEAs are non-invasive and allow real-time analysis of activity in multiple locations in cultured neurons. They are made up of a grid of tightly spaced electrodes which each record extracellular potentials and putative action potentials from surrounding cultured neurons. MEAs can also monitor network responses to drugs and see how cells communicate within culture, especially in response to drugs known to increase or decrease neural activity. Furthermore, newer systems allow measurements to be made from within an incubator, keeping cells as close to their optimum conditions for survival and hence, giving the most accurate results.
Where can we go from here?
The use of iPSC-derived neural cultures is becoming more widespread, with several industries using the technology as part of their drug screening process ( 5 ) . What remains to be seen is whether refinement of existing models can occur to the extent that the models are relevant to basic level human brain function and it is highly likely that co-culture of neurons with astrocytes will achieve this, opposed to neurons alone.
Moreover, the intricacy of the nervous system cannot be completely modelled in two-dimensions using only neurons and astrocytes. Indeed, microglia and oligodendrocytes have important roles within the CNS and the addition of these cell types adds further complexity and undoubtedly influences drug metabolism and responses to agents. Work to create these co-cultures in three-dimensions is under way , hoping to generate layers of the cortex as seen in vivo and provide a robust and relevant platform that not only looks like a human cortex, but behaves like one too.
1. Takahashi, K. & Yamanaka, S., 2006. Induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures by defined factors. Cell , Volume 126, pp. 663-676.
2. Chambers, SM. et al., 2009. Highly efficient neural conversion of human ES and iPS cells by dual inhibition of SMAD signalling. Nature Biotechnology , Volume 27(3), pp. 275-280.
3. Shi, Y., Kirwan, P. & Livesey, F., 2012b. Directed differentiation of human pluripotent stem cells to cerebral cortex neurons and neural networks. Nature , 7(10), pp. 1836-1846.
4. Odawara, A. et al., 2014. Long-term electrophysiological activity and pharmacological response of a human induced pluripotent stem cell-derived neuron and astrocyte co-culture. Biochemical and Biophysical Research Communications , Volume 443, pp. 1176-1181.
5. Authier, S. et al., 2016. Safety pharmacology investigations on the nervous system: an industry survey. Journal of Pharmacological and Toxicological Methods , Volume 81, pp. 37-46.