The expression of nestin in undifferentiated C17.2 cells is consistent with the presence of this intermediate filament protein in stem and progenitor cells in the mammalian CNS . However, as noted above, nestin mRNA was also readily detected in cells exhibiting morphological changes characteristic of differentiation, after one week in culture. Similarly, mRNA for the early neuronal marker, β-tubulin III, was found under all conditions examined, whereas GFAP mRNA was detected only in some cultures. These observations suggest that the C17.2 cells examined in this study are an heterogeneous population of stem and progenitor cells in keeping with evidence that NSCs exhibit morphological and phenotypic heterogeneity [12, 13].
The expression of diverse neurotrophins by NSCs is consistent with the role of these factors in the differentiation and development of the CNS. Presumably, the robust mRNA expression observed, particularly in cells maintained in 10% FBS + 5% HS, is driven by the serum-enriched milieu of potential inducers including neurotransmitters, hormones and growth factors, such as basic fibroblast growth factor and epidermal growth factor, which can stimulate C17.2 cell growth in vitro . In contrast to BDNF and NGF, which exhibited strong mRNA expression under all conditions examined, GDNF expression was weaker or not detectable in differentiating cells after seven days in culture. The suppression of GDNF expression might have been due to the prolonged exposure of NSCs to regulatory factors in the serum, as its decline appears to be inversely correlated with the concentration or enrichment of serum used for cell culture. Thus, moderate, weak or no expression of GDNF was observed in cells cultured for 1 week in 1% CS, 1% FBS or 10% FBS + 5% HS, respectively (see Fig. 1D,1F,1H). Various biological agents or pathways have been implicated in the regulation of GDNF expression. For example, fibroblast growth factor-2 and proinflammatory cytokines such as interleukin(IL)-1β, IL-6 and tumor necrosis factor-α stimulate GDNF synthesis and secretion . Activation of protein kinase C by phorbol esters increases GDNF expression [15, 16], whereas the adenylate cyclase activator, forskolin, inhibits GDNF production in cultured cells, suggesting an inhibitory role for the cyclic AMP- protein kinase A pathway . The cAMP pathway and its transcriptional factor cAMP response element binding protein (CREB) have been shown to induce differentiation in neuronal progenitor cells [17, 18]. Therefore, it is possible that activation of this pathway was involved in both the initiation of differentiation and the inhibition of GDNF expression observed in C17.2 cells after seven days.
While this work was in progress, it was reported that C17.2 neural stem cells constitutively secrete BDNF, GDNF and NGF, but do not label for GFAP or neuronal markers like β-tubulin III . Our findings are in agreement with these observations with regard to neurotrophin expression. However, in contrast to their findings, β-tubulin III mRNA and immunoreactivity were readily detected in our study. In addition, although GFAP mRNA was weakly expressed or not detectable in some cultures, immunoreactivity for this glial cell marker is present in C17.2 cells, as shown in Figure 4. These differences may be due to our examination of β-tubulin III and GFAP expression in cells maintained for 2–12 days in culture, whereas their C17.2 cells were examined after 2–3 weeks . Other factors, such as our use of low serum concentrations, as compared with the enriched culture medium used by Lu et al. , may also be involved. The detection of melatonin MT1 receptor mRNA in C17.2 cells after 2 days but not after 7 days, presumably involves downregulation of this receptor. There is considerable evidence that many G protein-coupled receptors are downregulated by their agonists . More importantly, melatonin, which is present in serum, has been found to suppress MT1 transcription in vitro . Interestingly, our immunocytochemical studies revealed MT1 immunoreactivity within C17.2 cell bodies and extensions, as shown in Figure 3A,3B. Although an intracellular localization could result from internalization of receptors , it is also possible that the immunoreactivity detected within these neural stem/progenitor cells is due to the presence of newly synthesized MT1 receptors. In accordance with this view, the MT1 protein detected in short-term (2-day) cultures is about 30 kDa, which is less than the approximately 37–45 kDa molecular weight observed in various mammalian tissues [22–24]. Moreover, when cells were cultured for 10–12 days, a MT1 receptor of about 40–45 kDa was detected, as shown in Fig. 2D. The mammalian MT1 contains two glycosylation sites in its N-terminal  and it may exist in more than one glycosylated form, as has been reported for other G protein-coupled receptors [25, 26]. Thus, the above cytochemical observations suggest that newly synthesized immature MT1 receptors, which have yet to undergo posttranslational modification and translocation to the plasma membrane, were detected in cells cultured for 2 days in 1% FBS, whereas a mature glycosylated receptor was present in cells grown for longer periods. Although the MT2 receptor transcript was not detected under any of the conditions used in this study, additional studies are required before the possibility of its expression in these cells can be ruled out. It is possible that MT2 mRNA may undergo rapid turnover/degradation, while a functional protein may still be present. This is the first evidence that melatonin receptors are expressed in neural stem or progenitor cells and raises the obvious question of whether this hormone plays a role in neuronal development. Although studies in this field are limited, there is increasing evidence that melatonin is involved in the early development of vertebrates. For example, melatonin is produced in chick embryos as early as the 7th day of embryonic development , and a physiological concentration of this hormone has been shown to significantly enhance mouse embryogenesis in vitro . Similarly, when sheep blastocysts were treated with melatonin for 24 hr in vitro, there was a significant increase in the percentage of embryonic survival . Other studies have shown that functional Gi protein-coupled melatonin receptors, which mediate inhibition of the adenylate-cyclase-cAMP pathway, are present in the embryonic (day 8) neural retina . Melatonin receptor transcripts for all the known Gi protein-coupled receptor subtypes have been found in 24 hr-old embryos from Japanese quail . Various studies have detected melatonin receptors in human fetal brain [31, 32] and peripheral tissues . Moreover, recent autoradiographic and in situ hybridization studies indicate that the melatonin MT1 receptor is expressed in diverse areas of the human fetal brain . Thus, the presence of MT1 receptors in NSCs is in keeping with the foregoing, and supports the view that melatonin is involved in neurodevelopment. Colocalization evidence that the MT1 receptor is present in both neural and glial progenitor cells is consistent with a neurodevelopmental role for melatonin, and suggests that in addition to the presence of the MT1 in mammalian neurons , it may also be expressed in astrocytes, as observed in similar cells from rat  and chick brain . The detection of nestin in some cells expressing the MT1 receptor is consistent with its presence not only in neural progenitor cells but also in GFAP positive glial progenitors . Preliminary evidence that melatonin induces GDNF mRNA expression in C17.2 NSCs, as we have observed previously in C6 glioma cells , supports the foregoing as this neurotrophic factor plays a critical role in both central and peripheral neurodevelopment [37, 38]. GDNF also exerts neuroprotective effects in the CNS, including a potent role in the survival of dopaminergic neurons in the midbrain [39, 40]. Therefore, modulation of GDNF expression may be one of the mechanisms underlying physiological neuroprotection by melatonin in the CNS .