During development, neurons undergo specific steps of maturation, including neurogenesis, neuronal migration, differentiation, and pruning. The cerebral cortex of vertebrates originates from the two telencephalic vesicles. The earliest neurons generated in the neuroepithelium of this vesicles form the PPL (Angevine and Sidman, 1961). The proliferative zones close to the ventricular surface, termed VZ and SVZ, subsequently continue to generate neurons, which migrate along radial glial processes toward the pial surface and split the PPL into the superficial MZ and the underlying SP. In addition, GABAergic neurons generated in the ganglionic eminences reach the immature neocortex by tangential migration (for review Marín and Rubenstein, 2001). Both MZ and SP are populated by specific neuronal populations, called CRc and SP neurons, respectively (Luhmann et al., 2009; Kirischuk et al., 2014). In between these two transient layers, the CP establishes, which will later develop into layers II to VI of the neocortex (see Bystron et al., 2008 for review). CRc in the MZ play an essential role for cortical lamination and disappear later during development, generating the cell sparse layer I (for review Kirischuk et al., 2014). SP neurons guide the projection of thalamocortical afferents and are transient synaptic targets of these axons before their final targets in layer IV appear (for review Kanold and Luhmann, 2010). The different neuronal populations that build up the mature neocortex later differentiate to form the right connectivity. Subsequently not only a considerable amount of neurons, but also ineffective connections are removed (for review Nikolic et al., 2013). A variety of reports demonstrated that neuronal activity and various neurotransmitters influence cortical development at different levels, ranging from early events as neurogenesis or migration to the establishment of fine-scale neuronal connectivity (for reviews Owens and Kriegstein, 2002; Spitzer, 2006; Wang and Kriegstein, 2009; Luhmann et al., 2016; Khazipov and Milh, 2017).

Immature neuronal activity starts with spontaneous calcium waves that have been observed in mouse neocortical slices already at early embryonic stages in both the VZ (Owens et al., 2000) and the early CP (Corlew et al., 2004). This early in vitro activity was replaced by a distinct set of large-scale network events, beginning with spontaneous and glutamate receptor dependent early network oscillations (Garaschuk et al., 2000), which were followed by giant depolarizing potentials a few days later, characterized by their dependence on depolarizing GABA_A receptor-mediated transmission (Allène et al., 2008). In vivo recordings confirmed that spontaneous discontinuous activity occurs already in the early postnatal neocortex (e.g., Khazipov et al., 2004; Hanganu et al., 2006; Yang et al., 2009; Colonnese and Khazipov, 2010). Already at birth local and short network oscillations, termed spindle bursts, in a frequency range of 10–20 Hz occurred, which were a few days later complemented by faster gamma oscillations with a frequency of 30–40 Hz (Khazipov et al., 2013; Yang et al., 2016). This discontinuous activity is generated in the sensory periphery, often independent of physical stimuli, subcortical regions, and in the immature neocortex itself (for review Luhmann et al., 2016; Khazipov and Milh, 2017).

These different patterns of neuronal activity play important roles for various physiological processes during neuronal development. Proliferation in the VZ is directly influenced by spontaneous calcium waves and spontaneous activity projected from the sensory periphery (Weissman et al., 2004; Bonetti and Surace, 2010). Spontaneous rhythmic intracellular calcium waves control neuronal migration (Komuro and Kumada, 2005). In addition, the rate of apoptosis as programmed cell death, which is an essential factor structuring the neocortical circuits, also directly depends on electrical activity (for review Blanquie et al., 2017a). And finally, spontaneous neuronal activity influences growth and differentiation of neuronal dendrites and axonal projections (for reviews Yamamoto and López-Bendito, 2012; Luhmann et al., 2016) and contributes to the formation of topographic maps (for reviews Hanganu-Opatz, 2010; Assali et al., 2014; Luhmann and Khazipov, 2017). While for many of these activity patterns, a vesicular release of GABA and glutamate is essential, recent studies identified taurine as a factor that contribute to GABAergic effects on neuronal development and also directly controls activity levels in the immature CNS as follows.

The neurogenesis in the VZ and SVZ is influenced by GABA_A receptors (LoTurco et al., 1995). Activation of GABA or glycine receptors also directly controls neuronal migration (for review Luhmann et al., 2015). Activation of GABA_A receptors induces apoptosis in CRc (Blanquie et al., 2017b), but promotes neuronal differentiation and synaptogenesis in principal cells and interneurons (Maric et al., 2001; Meier et al., 2003; see Wang and Kriegstein, 2009 for review). However, it should be noted in this respect that none of the mentioned studies unequivocally identified GABA as the endogenous neurotransmitter mediating these actions, but only demonstrated that GABA_A receptors are required. In fact, it has been shown that (i) neuronal migration is influenced by GABA_A receptors even in the absence of synaptic neurotransmitter release (Manent et al., 2005) and (ii) tonic GABA_A receptor mediated currents in embryonic neurons are unaffected by a massive GABA depletion, but sensitive to conditions enhancing extracellular taurine levels (Furukawa et al., 2014), allowing the speculation that taurine may contribute as endogenous agonist to the reported effects. With respect to the GABAergic and glycinergic actions, it must be considered that both neurotransmitters mediate a depolarizing action in the immature CNS, due to a high expression rate of the chloride loader NKCC1 and a low expression of the chloride extruder KCC2 (Yamada et al., 2004; Watanabe and Fukuda, 2015). Taurine is directly involved in this process, as it mediates an inhibition of KCC2 function via phosphorylation in immature neurons, thus maintaining a depolarizing action of GABA_A or glycine receptors (Inoue et al., 2012). In the following, we will summarize the action of taurine on distinct neuronal subpopulations and discuss how this interaction can interfere with neuronal development.

In the SVZ of the cortex, the activation of GABA_A receptors significantly reduced neurogenesis (LoTurco et al., 1995; Haydar et al., 2000). The observation that pharmacological inhibition of GABA_A receptors enhances the proliferation (LoTurco et al., 1995) indicates that an intrinsic agonist of GABA_A receptors controls neurogenesis. In the VZ, however, activation of GABA_A receptors increases proliferation (Haydar et al., 2000), probably by preventing neuroblast from exiting the proliferative cycle toward G0 phase. Since the proliferation of neuroblasts in the postnatal SVZ is limited by a nonsynaptic release of GABAergic agonists (Liu et al., 2005), it is intriguing to speculate that taurine contributes to these effects. Neuroepithelial stem cells directly contact cerebrospinal fluid (Lehtinen and Walsh, 2011; Lehtinen et al., 2011; Chau et al., 2015), which at least during early developmental stages subsequently to neural tube closing is consisting mainly of amniotic fluid, in which taurine is accumulated from maternal blood (Sturman et al., 1977; Sturman, 1981).

Recently it has been demonstrated that 10 mM taurine indeed increases the proliferation of mice embryonic progenitor cells in vitro (Hernandez-Benitez et al., 2010; Figure 3A). Similar effects were also found for human neuronal precursor cells, where proliferation is also enhanced by millimolar amounts of taurine, albeit it is not clear whether this process requires GABA_A or glycine receptors (Hernandez-Benitez et al., 2013). For hippocampal progenitors, it has also been demonstrated that 100 μM taurine augment proliferation (Shivaraj et al., 2012). In the mouse cochlea, millimolar taurine concentrations enhance proliferation of stem cells, but augment only the number of glutamatergic neurons, while the number of GABAergic neurons descending from cochlear stem cells is decreased (Wang et al., 2015). These results suggest that taurine can also provide a signal that determines the composition of neuronal elements within a circuitry. A similar observation was also made in the retina, where 100 μM taurine acting via glycine receptors stimulate the differentiation of retinal progenitor cells toward rod photoreceptors (Altshuler et al., 1993; Young and Cepko, 2004).

Intriguingly, electrophysiological experiments failed to reveal taurine-induced membrane currents in cells of the VZ (Flint et al., 1998), suggesting that these cells probably did not express taurine-sensitive glycine receptors. However, as from this publication, it is not clear whether the taurine experiments were performed in the presence of GABAergic antagonists, taurine probably affect neuroblasts exclusively via GABA_A receptors functionally expressed in VZ neuroblasts (LoTurco et al., 1995; Ma and Barker, 1995; Tochitani et al., 2010). On the other hand, it should also be considered that activation of GABA_A and/or glycine receptors can also influence proliferation indirectly, e.g., via interactions with bFGF or brain-derived neurotrophic factor (BDNF) release (Berninger et al., 1995; Antonopoulos et al., 1997).

Interestingly, taurine can also enhance adult neurogenesis under both in vitro and in vivo conditions (Hernandez-Benitez et al., 2012; Ramos-Mandujano et al., 2014; Gebara et al., 2015). This enhanced adult neurogenesis may underlie the beneficial effect of chronic taurine administration on various learning paradigms in adult mice (El Idrissi, 2008a; Neuwirth et al., 2013; Kim et al., 2014), although it should be noted that taurine can also directly affect in vitro correlates of memory formation (Chepkova et al., 2002; del Olmo et al., 2003; Sergeeva et al., 2003).

Already Sturman et al. (1985) suggested that in taurine deficient kitten neuronal migration in the cerebellum was hampered. In the visual cortex of these animals, clear indications for massively impaired migration were observed (Palackal et al., 1986). Taurine induced small inward currents in putatively migrating neurons of the rat cortical IZ at E19 via an activation of glycine receptors (Flint et al., 1998), whereas the taurinergic currents in clearly identified radially migrating neurons of the mouse neocortex were exclusively mediated via GABA_A receptors (Furukawa et al., 2014). In addition, it has been shown that GABA_A receptors (Barker et al., 1998; Heck et al., 2007), GABA_B receptors (Behar et al., 2001), GABA_C receptors (Denter et al., 2010), and glycine receptors (Nimmervoll et al., 2011), all possible targets of taurine, affect neuronal migration. Regarding the GABA_B receptors, Behar et al. (2001) directly demonstrated that taurine acts as potent chemoatractant for migrating neurons via an interaction with GABA_B receptors.

In dissociated cultures of cerebellar granule cells, taurine depletion indeed attenuated neuronal migration (Maar et al., 1995), which replicates the in vivo observation of Sturman et al. (1985) in the cerebellum of taurine deficient kitten. Recent in vivo studies by Furukawa et al. (2014) utilized GABA-deficient mice to specify the role of taurine for radial migration. Interestingly, they observed that radial migration was not significantly affected in a homozygous GAD-67 mouse, in which the GABA content was reduced to 12.7% of the wild-type level. However, inhibition of GABA_A receptors accelerated radial migration in this GAD-67 deficient mouse to a similar extent as in wild-type animals (Furukawa et al., 2014; Figure 3B), suggesting that GABA is not required as endogenous ligand for the GABA_A receptors regulating migration. This suggestion was substantiated by the findings that (i) the tonic currents in migrating neurons were unaffected by GABA depletion but (ii) enhanced in the presence of the TauT blocker GES, and (iii) attenuated after blockade of taurine synthesis with D-cystein (Furukawa et al., 2014; Figure 3B). And in line with these in vitro experiments, indicating that taurine is probably the most relevant endogenous agonist of extrasynaptic GABA_A receptors, depletion of taurine by maternal D-cystein administration accelerated migration in both wild-type and homozygous GAD-67 mice in vivo (Furukawa et al., 2014). In summary, these experiments provide mechanistic evidences that taurine is indeed a major endogenous modulator of radial migration.

In CRc, taurine-induced inward currents were mediated by glycine receptors with a low affinity (EC₅₀ = 2.4 mM, Kilb et al., 2002; Figure 2B). In line with the high intracellular Cl^- concentration in these cells (Achilles et al., 2007), activation of glycine receptors mediate a membrane depolarization in CRc (Kilb et al., 2002). Qian et al. (2014) were able to demonstrate that a strong electrical stimulation indeed stimulates the release of taurine and GABA, while glycine was not released. Voltage sensitive dye imaging revealed that such strong electrical stimulation induced a rapidly propagating wave of depolarization in the MZ of tangential slices. This propagating activity was insensitive to glutamate receptor blockade, but was partially attenuated by the application of either GABAergic or glycinergic antagonist, while application of both GABAergic and glycinergic antagonists completely abolished this activity (Qian et al., 2014). Because microdialysis experiments performed in this study demonstrate that glycine was not released (Qian et al., 2014), these results suggest a contribution of taurine. Since this propagating activity was observed at the pial surface of tangential slices in the MZ, most probably CRc contribute to this activity. Accordingly, whole-cell recordings from CRc revealed that these electrical stimuli induced a strong inward current that was also insensitive to glutamatergic antagonists, only partially reduced by GABAergic and glycinergic antagonists and completely blocked if GABAergic and glycinergic antagonists were combined (Qian et al., 2014). Since taurine was not localized in the presynaptic structures and glycine was not released after electrical stimulation (Qian et al., 2014), these results suggest that, although a substantial portion of this tangentially propagating activity transients is mediated by synaptic release of GABA, taurine acting on GABA_A and glycine receptors in CRc also considerably contributed to the propagation of activity in the MZ in a neuromodulatory fashion.

Taurine activates cortical neurons via glycine receptors at a rather low affinity with an EC₅₀ of about 1 mM (Flint et al., 1998; Okabe et al., 2004; Figure 2B). While no evidences for a synaptic activation of glycine receptor were found, inhibition of taurine transport with GES, hypoosmotic stimuli, and strong electrical stimulation evoke a putative nonsynaptic release of taurine, which activated glycine and probably also GABA_A receptors (Flint et al., 1998; Furukawa et al., 2014; Qian et al., 2014). Also in cortical neurons, the activation of glycine receptors mediates an excitatory effect, as concluded from taurine-induced Ca²⁺ transients and the fact that focal glycine application enhances the frequency of GABAergic PSCs (Flint et al., 1998). Sava et al. (2014) were able to replicate these findings and found that a prolonged application of 300 μM taurine induced a tonic inward current in putative projection neurons of the CP, which was in about 50% of the cells associated with a massive increase in GABAergic PSCs (Figure 4A). Pharmacological experiments revealed that about 80% of the tonic taurine-induced currents in these neurons are mediated via glycine, and about 20% via GABA_A receptors (Sava et al., 2014). The taurine-induced GABAergic responses are excitatory (Figure 4B) and are abolished in the presence of the glycinergic antagonist strychnine. As similar GABAergic PSCs could also be evoked by tonic application of glycine (Sava et al., 2014), it is concluded that glycine receptors on the presynaptic cells are essential to generate the GABAergic PSCs. Further analysis of the presynaptic GABAergic interneurons, which were visually identified in GAD67-GFP transgenic animals (Tamamaki et al., 2003), demonstrated that taurine induces a strychnine-sensitive inward current (Figure 4C), which was associated with a massive increase in action potential discharges in the majority of GABAergic interneurons (Figure 4D; Sava et al., 2014). This observation demonstrates that taurine indeed excites GABAergic interneurons. Further experiments revealed that these taurine-induced GABAergic PSCs increase the frequency of action potential discharges in putative pyramidal neurons (Sava et al., 2014). In summary, these studies demonstrated that taurine has an excitatory net effect in immature neuronal circuits in situ and thus most probably contribute network activity in the developing neocortex. However, it will be necessary to determine whether taurine contributes to excitation also in vivo, since recent experiments demonstrated depolarizing, but inhibitory effects of GABA_A receptors in the developing neocortex under in vivo conditions (Kirmse et al., 2015).

For SP neurons, it was also demonstrated that taurine activates glycine receptors with a low affinity (EC₅₀ = 1.7 mM; Kilb et al., 2008). While higher taurine concentrations evoke desensitizing responses, 100 μM taurine induced a tonic inward current in this cell type (Figure 2B). Also in SP neurons, the activation of glycine receptors mediates a depolarizing action (Kilb et al., 2008). Further analyses revealed that tonic taurinergic currents, although evoking only small subthreshold depolarizations, substantially lower the action potential threshold, thus demonstrating an obvious excitatory effect on SP neurons. Inhibition of the taurine transport with GES as well as hypotonic stimulation induced inward currents, which suggests that also in the SP neurons taurine can be an endogenous agonist of glycine receptors (Kilb et al., 2008). In summary, these experiments indicate that taurine can act as neuromodulator in the SP. As a fraction of SP neurons are GABAergic, it was also investigated whether SP neurons contribute to the taurine-induced GABAergic PSCs in pyramidal neurons (see section “Neurons in the Cortical Plate or Developing Cortical Layers”). However, these experiments revealed that ablation of the SP does not significantly reduce the frequency of taurine-induced GABAergic PSCs, demonstrating that GABAergic projections from SP to pyramidal neurons do not considerably contribute to the taurine-induced activity in the CP (Sava et al., 2014). On the other hand, ambient taurine level could be highest in the embryonic SP (Furukawa et al., 2014). Since the SP is a crucial element for structural and functional development of the neocortex (for review Kanold and Luhmann, 2010) and is essentially involved in the regulation of excitability in the developing neocortex (Dupont et al., 2006; Hanganu et al., 2009), further experiments are needed to evaluate whether SP neurons are the particularly important target for the neurodevelopmental effects of taurine.