Neuronal migration is a tightly regulated process in which the orchestrated activity of multiple molecules combines to bring about the appropriate allocation of post-mitotic nerve cells. Previous studies have shown that the NRG1/ErbB4 system is involved in several types of neuronal migration. On the other hand, intracellular calcium fluctuations, in many cases regulated by influx through Ca2+ permeable membrane channels, play central roles in neuronal motility [32, 34]; however, little is known about the involvement of calcium signals in NRG1-induced neuronal migration. In this work, we analyzed the calcium dependence of the migratory-promoting activity of the NRG1-ErbB4 system in immortalized neural progenitors (ST14A cells), and the interaction between this pathway and the one dependent on the NMDA-type glutamate receptor, a calcium permeable channel. Using ST14A cells exogenously expressing ErbB4, we demonstrate that the two systems synergistically act in stimulating progenitor cell migration in a Ca2+-dependent manner.
ST14A cells were chosen because they endogenously express all members of the ErbB family (ErbB1, ErbB2, and ErbB3) except ErbB4. We were able to compare migratory activity in cells lacking and expressing ErbB4 by using wild type ST14A cells and clones permanently transfected with specific ErbB4 isoforms . Our RT-PCR studies show that ErbB4-transfected cells, similarly to wild type ST14A cells, express a restricted number of NMDAR subunits: NR1.4a, NR2C, NR2D and NR3B. Data from the literature indicate that the NR1 subunit is essential for NMDAR trafficking and insertion into the plasma membrane; different combinations of the NR2 and NR3 subunit families generate NMDA receptors differing in their functional characteristics, and cell-type and sub-cellular localization . The two NR2 subunits expressed by ST14A cells (NR2C and NR2D) are prevalently extra-synaptic and are characterized by low-conductance openings and low sensitivity to Mg2+ block [35–37]. On the other hand, NR3 subunits act like a dominant negative, reducing the amplitude of the response and its desensitization, and decreasing the Ca2+ permeability of the NMDAR channel . Since usually NMDARs are composed of at least one NR1 subunit and variable combinations of NR2/NR3 subunits, the resulting NMDARs would generate smaller, long lasting currents, that are less Mg2+-sensitive and Ca2+-permeable than those of NR2A-2B containing NMDARs.
We did not find a statistically significant effect of NMDA on basal migration of either wild type or ErbB4-transfected ST14A cells; this finding could be ascribed either to a low number of NMDA receptors on the ST14A cell membrane, or to their reduced Ca2+ permeability, yielding a Ca2+ influx not sufficient to elicit a migratory response by itself.
ErbB4 expression is necessary for NRG1-induced migration in ST14A progenitors. The mechanism has been partly characterized in our laboratory in previous studies : ErbB4 dimerization and tyrosine kinase trans-phosphorylation is followed by intracellular signalling cascades involving PI3K activation. When NMDA was added to NRG1 in migration assays, we found a significant enhancement of migration compared to NRG1 alone (NRG1 alone: 5 fold, NRG1 + NMDA: 8 fold increase in migration rate). This strengthening of NRG1-stimulated migration is suggestive of a cooperation between NMDA and ErbB4 receptors in inducing and/or maintaining a migratory behaviour in ST14A cells. The NMDA-mediated increase in NRG1-stimulated migration could be prevented by the addition of a NR2C/D subunit-specific antagonist (UBP141), strongly suggesting that the NMDARs cooperating with ErbB4 during ST14A cells migration contain NR2C/D subunits.
The role of increases of the cytosolic calcium concentration in controlling cell motility has been reported in various cell types ranging from fibroblasts to immature neurons [39, 40]. Neuronal precursors and post-migratory neurons in the fetal cerebrum and the early postnatal cerebellum exhibit spontaneous Ca2+ transients [41, 42], suggesting that the role of Ca2+ transients in development may be constantly remodelled by internal programs and extracellular cues. Ca2+ transients may affect the recycling of cell-adhesion receptors and induce the rearrangement of cytoskeletal components, which are essential for cell movement. The spontaneous Ca2+ transients in migrating granule cells are mediated by NMDA receptors and by N-type Ca2+ VDCCs [10, 34]. Other agonists that can trigger an elevation in [Ca2+]i involved in motility control are the neurotransmitter GABA, that exerts an excitatory action in immature cells due to the high intracellular Cl- concentration, the neuropeptide somatostatin that influences Ca2+ fluctuations in a stage-dependent manner, and neurotrophins, like Brain Derived Neurotrophic Factor [13, 43–45]. The ErbB4 receptor, like other members of the ErbB family [46, 47], upon ligand activation could elicit calcium release from intracellular stores and calcium influx. However, there are no data available for this mechanism in neuronal cells, apart form a report  indicating that NRG1 can modulate glutamate-induced [Ca2+]i increase. Therefore, we checked if the synergism of NMDAR and NRG1/ErbB4 systems in enhancing ST14A migration could be due to a modulation of Ca2+ signalling.
In migration experiments carried out in the presence of the intracellular Ca2+ chelator BAPTA-AM we found a highly significant reduction both in NRG1 and in NMDA+NRG1 stimulated migration, with the latter reduced to the same level seen with NRG1 alone. This result strongly supports the hypothesis that both NRG1 induced migration in ST14A cells and the synergistic effect of NMDAR/ErbB4 activation involve intracellular Ca2+ changes. For NRG1-induced migration, values in the presence of BAPTA-AM were still significantly higher than controls, indicating that NRG1/ErbB4-induced migration could involve also Ca2+-independent mechanisms.
In the attempt to understand the mechanisms underlying these calcium-dependent effects, we began to characterize the Ca2+ signalling induced by NRG1 and NMDA by means of calcium imaging experiments. Our findings indicate that NRG1 stimulation of ErbB4-transfected ST14A cells induces a long lasting increase in [Ca2+]i, while no response could be recorded from mock transfected cells. To clarify the relative contribution and extent of calcium influx from the extracellular medium and of release from internal stores, NRG1 stimulation was repeated in a medium without Ca2+ and containing the calcium chelator EGTA. Cells responded to NRG1 with an increase in free calcium, but the signal was significantly smaller and showed a transient time course. Intriguingly, after termination of NRG1 stimulation and subsequent reinstatement of Ca2+ in the medium, the level of [Ca2+]i was strongly increased. This finding can be explained by the activation of a store-operated calcium entry (SOCE) driven by channels activated in response to depletion of the intracellular Ca2+ stores (see e.g. 33). Accordingly, when the mechanism was activated by emptying the intracellular stores by means of thapsigargin, cells did not respond to a subsequent stimulation with NRG1.
Stimulation with NMDA led to an increase in free intracellular Ca2+ that was significantly smaller than the one obtained with NRG1, could be observed in only 46% of cells tested and was completely abolished when the cells were preincubated either in a calcium-free solution or in the presence of the NR2C/D subunit-specific NMDAR antagonist UBP141. Significantly, these signals showed a very slow time course: more than 10 min were required to obtain a plateau value. This finding can be explained by a very low level of activation of NMDA channels at resting potentials, compatible with the proposed reduced Mg2+ sensitivity of the subunits expressed in our cells; a limited activation of these channels could induce some depolarization, thus slowly recruiting additional channels. The slow time course of activation of this mechanism, and the relatively low global increase in calcium levels could not be sufficient to promote cell motility.
When both agonists were applied simultaneously, or when NRG1 was presented during the slow rising phase of the NMDA response, peak amplitudes were not significantly different from those obtained following stimulation with NRG1 alone; however, when NMDA was removed during the plateau phase of the response, that can be ascribed to calcium influx, a slight reduction in [Ca2+]i could be detected. Taken together, these observations can explain the different roles of the two agonists in promoting migration of ST14A cells and their calcium dependence: the slow and small increases in [Ca2+]i elicited by NMDAR activation are not sufficient to promote a significant amount of neuronal migration, while the faster, stronger and sustained signals induced by the NRG1/ErbB4 pathway are sufficient; however, when NMDA is presented together with NRG1, the small additional amount of calcium entering from the extracellular medium can exert a synergistic effect, possibly by inducing calcium signals localized to cellular subdomains (e.g. submembrane regions) that could contribute to the enhancement of cell motility.
NRG1 is one among several candidate genes linked to a relevant behavioural disease, schizophrenia. There are many hypotheses on the pathogenesis of this disease: some studies suggest that young-adult onset of schizophrenia may originate from disarranged circuitry caused by an altered migration of neuronal precursors during pre- and post-natal brain development [49, 50]. According to post-mortem studies, in the brain of schizophrenic patients there is an altered NRG1/ErbB4 signalling and an hypo-functional glutamatergic system. Therefore, the understanding of NMDAR-NRG1/ErbB4 interactions during neural precursor migration may contribute to reveal the molecular and cellular basis of schizophrenia and of other neurodevelopmental disorders.