More than one way to induce a neuron

Seventy-six pairs of transcription factors can induce mouse connective-tissue cells to adopt a neuron-like identity in vitro. This discovery provides insights into both neuronal development and cell reprogramming.
Lynette Lim is in the Centre for Developmental Neurobiology, Institute of Psychiatry, Psychology and Neuroscience, King’s College London, London SE1 1UL, UK, and in the MRC Centre for Neurodevelopmental Disorders, King’s College London.

Search for this author in:

Oscar Marín is in the Centre for Developmental Neurobiology, Institute of Psychiatry, Psychology and Neuroscience, King’s College London, London SE1 1UL, UK, and in the MRC Centre for Neurodevelopmental Disorders, King’s College London.

Search for this author in:

The brain contains hundreds of neuronal subtypes, each defined by a specific combination of features, including its position and shape, the neurotransmitter molecules it produces and its electrophysiological properties1. Engineering this enormous diversity in the laboratory is an ultimate goal of regenerative medicine. In a paper in Nature, Tsunemoto et al.2 describe a large-scale effort to identify factors that can endow non-neural cells cultured in vitro with neuronal properties.

Understanding the mechanisms that underlie the generation of neuronal diversity has been a central goal of neurobiology for more than a hundred years, since the neuroscientist Santiago Ramón y Cajal postulated that the nervous system is made up of discrete individual cells3. Work over the past three decades has identified gene regulatory networks that control neuronal identity as it unfolds in the embryonic brain4. These studies have also revealed that neuronal identity is intimately linked to the environment in which neurons develop, primarily because some of the cells’ most important features, such as their connections, depend on their interactions with other neurons.

In the past decade, however, it has become clear that many neuronal attributes can be generated outside the normal context of brain development. For example, in 2010 it emerged5 that a cocktail of three transcription factors can be applied to fibroblasts (the most common cells of connective tissue) cultured in vitro to convert them into cells that resemble brain-derived neurons, at least in terms of their shape and electrophysiological properties. This procedure, called direct lineage reprogramming, is based on the premise that certain transcription factors regulate gene-expression patterns characteristic of neuronal cell types. But what has not been clear is whether the capacity to reprogram cells into neurons is limited to a handful of transcription factors.

Previous work by the group that performed the current study showed that a pair of transcription factors from the basic helix-loop-helix (bHLH) and Pit-Oct-Unc (POU) families can induce the expression of neuronal markers through direct reprogramming6. Tsunemoto et al. were inspired by this finding in their current work. The authors screened 598 pairs of bHLH and POU transcription factors — chosen on the basis of their expression in neuronal lineages — to see which could transform mouse embryonic fibroblasts into neurons in vitro (Fig. 1). Seventy-six of the pairs produced cells that expressed multiple markers of mature neurons and had neuronal morphologies. Thus, neuronal features can be induced in non-neuronal cells by an astonishing range of transcription-factor combinations.

Figure 1 | Inducing neurons through direct reprogramming. Tsunemoto et al.2 grew embryonic fibroblasts (connective-tissue cells isolated from mouse embryos) in vitro. They treated the cells with 598 pairs of transcription factors from the basic helix-loop-helix (bHLH) and Pit-Oct-Unc (POU) families, most of which are expressed in neuronal lineages. In total, more than 12% of the transcription-factor combinations tested could reprogram the fibroblasts into cells with neuronal properties (indicated with ticks). Different pairs produced neuron-like cells with different shapes, gene-expression profiles and electrophysiological properties (indicated by different colours).

How similar are the neurons induced by the transcription-factor pairs? Analysis of gene expression using single-cell RNA sequencing revealed that a given pair of factors generates relatively homogeneous populations of neurons, which share a similar molecular profile. This is surprising, because previous experiments have highlighted the heterogeneity of cell populations undergoing reprogramming in culture7.

By contrast, Tsunemoto et al. found that different transcription-factor combinations induced the formation of distinct neuron-like populations that had characteristic markers and electrophysiological features. However, the authors also found that certain features — such as the expression of particular ion channels or neurotransmitter receptors — could be induced by multiple combinations of transcription factors. These findings support the idea that there is not a single ‘gene code’ for making a particular feature of neuronal identity, but that the molecular machinery underlying neuronal development has a remarkable degree of redundancy. In other words, identity-defining transcription factors are probably used in different combinations in distinct neuronal cell types to generate the diversity seen in vivo.

One key question is whether the induced neurons faithfully mimic neuronal cell types found in vivo, or whether (and to what extent) they represent artificial cell types. To address this issue, the authors compared gene-expression patterns in induced cell populations with those in neuronal subtypes taken from three-week-old mice. This analysis indicated that the induced neurons did not match endogenous cell types of the juvenile mouse brain. However, it might be that the induced neurons did not reach the same stage of development as endogenous cells. Matching cell types across different developmental stages remains a complex challenge in neurobiology, as shown this year by two RNA-sequencing studies8,9. These highlighted the difficulty of recognizing the transcription factors that define specific cell types at early stages of neuronal development in the brain’s cerebral cortex.

It is possible, then, that the neurons induced in vitro by Tsunemoto et al. correspond to specific populations of endogenous neurons in a relatively immature state. These cells might develop into fully differentiated neurons only if placed in the appropriate environment. Alternatively, the expression of a pair of transcription factors might be sufficient to elicit the development of some neuronal features in fibroblasts, but not to unleash the complete program of differentiation that takes place in the embryo.

Nevertheless, Tsunemoto and colleagues’ study adds to the body of evidence showing that some features of neuronal identity can be reproduced outside the developing brain. In doing so, it demonstrates the power of reprogramming to interrogate the function of neuron-specific genes. The authors have made their findings available in a database (see that will allow other researchers to use the transcription-factor codes to induce specific neuronal features. This will doubtless prove useful for studying the selective vulnerability of specific neuronal subtypes to disease.

Finally, the authors provide preliminary evidence that their transcription-factor combinations can also be used to generate neurons from human embryonic fibroblast-like cells. Following further validation, the codes might help us to decipher the origins of neuronal diversity in humans.

Nature 557, 316-317 (2018)

doi: 10.1038/d41586-018-04978-5
Nature Briefing

Sign up for the daily Nature Briefing email newsletter

Stay up to date with what matters in science and why, handpicked from Nature and other publications worldwide.

Sign Up


  1. 1.

    Zeng, H. & Sanes, J. R. Nature Rev. Neurosci. 18, 530–546 (2017).

  2. 2.

    Tsunemoto, R. et al. Nature 557, 375–380 (2018).

  3. 3.

    Ramón y Cajal, S. Histologie du Système Nerveux de l’Homme et des Vertébrés (Maloine, 1909).

  4. 4.

    Lodato, S. & Arlotta, P. Annu. Rev. Cell. Dev. Biol. 31, 699–720 (2015).

  5. 5.

    Vierbuchen, T. et al. Nature 463, 1035–1041 (2010).

  6. 6.

    Blanchard, J. W. et al. Nature Neurosci. 18, 25–35 (2015).

  7. 7.

    Treutlein, B. et al. Nature 534, 391–395 (2016).

  8. 8.

    Mi, D. et al. Science 360, 81–85 (2018).

  9. 9.

    Mayer, C. et al. Nature 555, 457–462 (2018).

Download references