Genetic Program of Neuronal Differentiation and Growth Induced by Specific Activation of NMDA Receptors
- First Online:
- Cite this article as:
- Ghiani, C.A., Beltran-Parrazal, L., Sforza, D.M. et al. Neurochem Res (2007) 32: 363. doi:10.1007/s11064-006-9213-9
Glutamate and its receptors are expressed very early during development and may play important roles in neurogenesis, synapse formation and brain wiring. The levels of glutamate and activity of its receptors can be influenced by exogenous factors, leading to neurodevelopmental disorders. To investigate the role of NMDA receptors on gene regulation in a neuronal model, we used primary neuronal cultures developed from embryonic rat cerebri in serum-free medium. Using Affymetrix Gene Arrays, we found that genes known to be involved in neuronal plasticity were differentially expressed 24 h after a brief activation of NMDA receptors. The upregulation of these genes was accompanied by a sustained induction of CREB phosphorylation, and an increase in synaptophysin immunoreactivity. We conclude that NMDA receptor activation elicits expression of genes whose downstream products are involved in the regulation of early phases of the process leading to synaptogenesis and its consolidation, at least in part through sustained CREB phosphorylation.
KeywordsN-methyl-D-AspartateCortical neuronsDNA arrayGene regulationCREBBrain development
Brain derived neurotrophic factor
cAMP responsive element binding
CREB binding protein
Immediate early genes
Long term potentiation
Neuronal activity-regulated pentraxin
Light molecular-weight neurofilament
NGF-induced factor B
Neuron-derived orphan receptor
Mammalian brain development is modulated by the orchestrated interaction of cell intrinsic and extrinsic factors that, in turn, can be influenced by environmental stimuli. In utero exposure to harmful substances connected with maternal behaviour and/or exposure (e.g. consumption of neurotoxic substances) may induce untimely interactions between the cell intrinsic and extrinsic signalling factors, interfering with proper brain development and increasing the risk of neurodevelopmental diseases [1–4].
Among the extrinsic factors, neurotransmitters and their receptors play key roles in central nervous system (CNS) development . Glutamate is the major excitatory neurotransmitter in the brain. Its receptors are expressed early in development by neural progenitors and by selected early neuronal populations. This suggests that they might play a role in directing the development of later populations, and in the proper wiring of the developing CNS [5–7]. Glutamatergic actions are mediated through 3 types of ionotropic (ligand-gated ion channels) receptors (iGluR: N-methyl-d-Aspartate (NMDA), α-Amino-3-hydroxy-5-methylisoxazole-4-propionic acid (AMPA) and Kainate) and one group of G-protein coupled metabotropic receptors (mGluR) of which several functional subunits have been identified [8, 9]. The expression of these receptor subtypes and their subunits appears to be developmentally regulated, as some of them are only expressed during embryonic or early postnatal life. In order to properly influence developmental events, each of these subtypes and some of their subunits need to be expressed at the right time. Changes in subunit composition drive their effects and are essential for proper brain development .
Among the iGluR, NMDA receptor activation regulates developmental programs during embryonic and perinatal periods. These receptors directly or indirectly influence cell proliferation, migration, cytoskeleton-related protein synthesis, and the formation and establishment of synapses [5–7, 10–12]. Most importantly, it was reported that blockage of NMDA receptors during brain development triggers neuronal cell death , further supporting the notion that in addition to support neuronal development and the development of neuronal circuits, they are important for neuronal survival. NMDA receptor activity is highly enhanced during perinatal life to promote activity-dependent synaptic connections. Immature NMDA receptor-gated channels open easily resulting in more calcium influx than in the adult brain . Therefore, any event or stimulus that interferes with this “normal” level of synaptic activity will disturb brain development and may lead to life-long disabilities. For example, abnormal stimulation of glutamatergic receptors or excessive levels of glutamate can lead to excitotoxicity and subsequent perinatal brain injury . Thus, it is extremely important to understand which target genes are involved in NMDA actions as well as which factors may modulate NMDA-mediated “good” or “bad” effects . In addition, knowledge on NMDA-induced gene regulation could help developing strategies for alleviating neurodevelopmental deficiencies.
In the adult brain [9, 15] NMDA receptors have been implicated in activation of signalling pathways leading to synaptic plasticity, process outgrowth and strengthening of pre-existing connections. Their activation regulates early and late phases of synaptic formation and consolidation. These cellular programs involve activation of transcription factors and immediate early genes (IEGs), up-regulation of gene expression and de novo protein synthesis [16–21]. The role of NMDA receptors on gene expression mediated by induction of IEGs and transcription factors has been extensively studied in the hippocampus [22–25], and after injury or stress [26–28] both in vivo and in vitro. However, a detailed analysis of the effects of NMDA receptor activation on regulation of gene expression in embryonic cortical neurons in the complete absence of glial cells is still lacking. Furthermore, little is known about the effects that a massive NMDA receptor activation can have on young neurons .
We have previously shown that massive activation of NMDA receptors in primary neuronal cultures from rat embryonic cerebral cortex elicited changes in the expression of IEGs and transcription factors such as: cfos, NGFI-B, Krox-24. These effects were accompanied by changes in cfos protein levels and sustained increase in intracellular calcium. NMDA even at the very high dose of 1 mM did not trigger apoptosis, which appeared to be mediated through AMPA receptors .
In the present report by using the same culture system  in which neuronal responses can be measured in the absence of glial cells, a cell composition resembling an immature brain, we show evidence for a link between NMDA receptor activation and gene regulation mediated, at least in part, through sustained induction of cAMP responsive element-binding (CREB) phosphorylation. Strong evidence of the involvement of the transcription factor CREB as mediator/effector of NMDA receptor action on gene expression are still lacking. Data presented in the literature remain controversial (for review: Platenik et al. 2000, ). We addressed the role of NMDA receptors in the program of cortical neuronal maturation, using our serum-free neuronal specific culture medium and gene array analysis. Our work shows that the use of the combination DNA microarray-Real-time RT-PCR is a reliable and quantitative tool to assess gene regulation in cortical neurons in vitro.
NMDA, creatine, poly-d-lysine were purchased from Sigma, Saint Louis, MO, USA. Human recombinant bFGF, B27 supplement and DMEM-F12 were from Invitrogen Life Technologies, Carlsbad, CA, USA. The antibody directed against phosphorylated Ser133 CREB (pCREB; rabbit polyclonal) was purchased from Upstate Biotechnology (Lake Placid, NY, USA). Anti-MAP2 (mouse monoclonal) and anti-synaptophysin (mouse monoclonal) were from Sigma (Saint Louis, MO, USA). Anti-GFAP (mouse monoclonal and rabbit polyclonal) was from Neomarkers (Fremont, CA, USA). All secondary Affinipure fluorochrome-conjugated antibodies were purchased from Jackson ImmunoResearch Laboratories Inc. (West Grove, PA, USA).
Primary rat neuronal cortical culture
Cultures were prepared as previously described [30, 32] with minor modifications. E16 Sprague-Dawley rats (Charles-River, Wilmington, MA, USA) were killed following the National Institutes of Health and UCLA animal welfare guidelines. Cerebral cortices were removed, combined in DMEM-F12 containing 10% pure fetal calf serum (Atlanta Biologicals, Norcross, GA, USA) and mechanically dissociated for 2 min using a Stomacher 80 (Seward, London, UK). The dissociated cells were filtered in turn through 140 μm and 230 μm sieves (Cellector, E-C Apparatus Corp., Holbrook, NY, USA) to remove large clusters. Sieves were washed sequentially with DMEM-F12 containing 10% pure fetal calf serum. Cells were collected by centrifugation, and cell pellet was resuspended in a neuronal specific medium , TII, supplemented with bFGF (10 ng ml-1), B27 (1:50), and creatine (2 mg ml−1). The cells were plated onto freshly poly-d-lysine coated dishes and cultured for 10–11 days unless differently stated. Purity of the cultures was assessed by double immunostaining with neuronal, astrocytic, oligodendrocyte and microglial markers. The cultures were 98% immunopositive for MAP2 and the remaining immunostained for astrocytic or oligodendrocyte markers. No microglial cells were detected.
DNA array analysis
Total RNA was extracted from primary neuronal cortical cultures grown on 100 mm dishes (1.5 × 10−6 cells ml−1) for 11 days by using TRIzol reagent (Invitrogen, Life Technologies, Carlsbad, CA, USA), cleaned with RNeasy (Qiagen, Valencia, CA, USA), and its quality was monitored by micro-capillary electrophoresis (Bioanalizer 2100, Agilent Technologies). Hybridisation and scanning of the Genechip Rat Neurobiology U34 oligonucleotide arrays (Affymetrix Inc., Santa Clara, CA, USA) was performed at the UCLA Microarray Core Facility following protocols recommended by the manufacturer. Expression values, representative of the amount of transcripts in solution, were calculated using Robust Microarray Analysis (RMA) . Hybridisation experiments were repeated either 3 or 5 times using independent RNA samples. Data represent the average of all independent hybridisations in each experimental condition. Before further analysis, we selected the genes whose expression was more than 2-fold or less than 0.5-fold when compared with the control. After this selection, a t statistic was used to determine which genes were differentially expressed. Differences in expression were considered significant when p < 0.05 for Student’s t test. Final list of interesting genes was obtained by filtering all genes that were significantly, differentially expressed in at least one condition. The software GENECLUSTER was used to perform the mathematical calculations and to obtain convenient data visualisation .
Oligonucleotide primers used for Real-time PCR
Gene (accession #)
Annealing temperature (°C)
AAG TCC ACG TGC TCC CAG T
β2-microglobulin (NM _012512)
Neuronal cultures were plated onto glass coverslips (125000 cells ml−1) coated with Poly-d-Lysine and cultured for 11 days. The cells were fixed in 4% paraformaldehyde in PBS for 15 min; permeabilised for 10 min in 0.2% Triton X−100 and blocked in 10% normal goat serum for 15 min; all these steps were performed at room temperature. Primary antibodies were diluted in PBS + 1.5% goat serum and incubated overnight at 4°C. The coverslips were washed thoroughly in PBS and incubated for 30 min at room temperature with the secondary antibodies. After extensive washes, the coverslips were mounted in Vectashield (Vector Laboratories, Burlingame, CA, USA). Pictures were taken using an Olympus fluorescent microscope with a Zeiss digital camera or a Zeiss confocal microscope (LSM 510 Meta systems).
Western blotting analysis
Following NMDA (1 mM)-treatment (see Figure legends for details), 11-day-old primary neuronal cortical cultures were washed and harvested in PBS. Western Blots were performed as previously reported [37, 38]. Equal protein loading was verified by Ponceau S solution reversible staining of the blots. Relative intensities of the protein bands were quantified by scanning densitometry using the public domain NIH Image program (NIH image, ImageJ, http://rsb.info.nih.gov/ij/). Statistical analysis was performed using GraphPad Prism 4.01 (GraphPad Software, Inc., San Diego, CA, USA).
Fluorescence imaging was performed using a custom-built video-rate confocal microscope, as previously described . In brief, cells on coverslips were loaded with the Ca2+ indicator fluo−4 by incubation in 3 μM fluo-4 AM for 20–30 min. Cells were excited with a 475-nm diode laser and imaged with 20× or 40× water immersion objectives. The resulting fluorescence was bandpass filtered at 510 nm and detected by a photomultiplier tube (Hamamatsu). Excitation and emission were scanned with two oscillating mirrors (GSI Lumonics). Images were captured with a Raven board (Bit Flow, Inc.) using Video Savant software. Maximum image resolution was 980X730 at 15 frames per second. Analysis of imaging data was performed using ImageJ software. Cells were considered to have a response to NMDA if they showed either a >20% increase in fluorescence over background, or at least a 50% increase in the frequency of spontaneous Ca2+ transients in response to bath application of NMDA.
Primary cortical neurons express NMDA receptor subunits
In the present work, we used primary cortical neuronal cultures from 16-day rat embryos. RNA extracted from primary cortical neurons at different days in culture was analysed by Real-time RT-PCR for the expression of subunits for NMDA receptors. We focus our attention on receptor subunits known to be highly represented in the cerebral cortex, and to be present since early development [8, 10].
Consistent with the receptor subunit expression, only a small percentage of cells responded to NMDA (10 μM) with an increase in intracellular Ca2+ at day 1. The Ca2+ response to NMDA increased with culture age up to 10 days, at which time it remained relatively constant through 14 days in culture (Fig. 1B). As previously reported , lower concentrations of NMDA (1 or 5 μM) resulted in an increase in the frequency of spontaneous Ca2+ transients, whereas higher concentrations of NMDA (>10 μM) evoked a sustained increase in [Ca2+]i in all responding cells.
NMDA induces time-dependent changes in gene expression in cortical neurons
Our neuronal culture system resembles the immature brain before glial development occurs. This model has allowed us to investigate glutamatergic and nicotinic actions on primary neurons, without inducing a response in other types of cells, such as oligodendrocytes and astrocytes that are known to express functional neurotransmitter receptors. Using this model, we previously reported  that treatment of cortical neurons with 1 mM NMDA for increasing periods of time (10–360 min) triggered changes in IEGs or transcriptions factors involved in the control of gene expression without inducing cell death, which appeared to be selectively elicited by activation of the glutamatergic receptor subtype AMPA.
Brief activation of NMDA receptors induces delayed changes in gene expression
Since short treatments with NMDA (1 mM) triggered activation of IEGs and transcription factors , it was important to investigate any possible downstream effect that activation of these IEGs and transcription factors could have on other genes.
We found upregulation in the expression of IEGs and transcription factors. The IEG t-PA, reported to be induced during LTP, seizures and kindling  was increased 4.0-fold by a 10 min NMDA-treatment. Likewise, NGFI-B and c-fos (3.2 and 3.9-fold, respectively) were significantly upregulated in the samples that had been treated for 10 min with NMDA together with NOR-1 and ania-3 (both 2.7-fold over control-untreated cortical neurons, Fig. 3, upper panels). Ania-3 expression was not significantly affected by NMDA activation for 30 and 60 min (1.8 and 1.9-fold over control), whereas upregulation of other IEGs/transcription factors was sustained (c-fos: 3.3 and 2.9-fold, respectively for the other two treatments; NGFI-B: 2.8-fold at both times; and NOR-1: 2.4 and 2.1-fold). The murine osteosarcoma provirus (m.o.p.) expression (3.0-fold) was higher in the samples treated for longer time (30 min) together with another probe set for cfos (2.9-fold).
Only two genes were significantly downregulated (<0.5-fold; Fig. 3, upper left panel) neural membrane protein 35 (NMP35) or Lifeguard and the metabotropic glutamate receptor 3 (mGluR3). The different probes for the same gene in the microarrays showed similar pattern (Fig. 3, upper panels).
Nine genes were then, selected for verification by Real-time RT-PCR (Fig. 3, bottom panel). We found a complete agreement between the changes in gene expression found with DNA arrays and Real-time RT-PCR quantification.
NMDA-induced changes in gene expression are concentration-dependent
To exclude that the changes in gene expression could be due to activation of other glutamate receptor subtypes by a very high concentration of NMDA and to demonstrate that NMDA receptors were active also at lower concentrations and there was no need to overcome a blockage of these receptors due to the magnesium present in the medium, we measured the effect of increasing concentrations of NMDA on BDNF gene expression by Real-time RT-PCR, 24 h after treatment.
NMDA-induced CREB phosphorylation is long-lasting
Change in gene expression is accompanied by an increase in synaptophysin immunoreactivity
The question of how neural progenitors and immature neurons respond to the excess of glutamate released after an insult remains unresolved. In the present study, we focused on the role of NMDA receptors in the program of cortical development, using our serum-free neuronal specific culture medium and gene array analysis. An advantage of our culture system was to be able to study the effect of NMDA receptor activation in the absence of other neuromodulators released from surrounding cells, principally microglia and astrocytes, which can contribute to the excitotoxicity of glutamate and NMDA.
In an attempt to study the response(s) of young and still developing neurons to a major insult, we used a high concentration of NMDA (1 mM). Interestingly, this treatment did not result in massive cell damage followed by apoptosis . Hence, we excluded a role for NMDA receptor activation in cell death and predicted that a considerable activation of NMDA glutamatergic receptors could be, on the other hand, beneficial for the developing neurons. Our contention is supported by previous studies in which blockage of NMDA receptors during development triggered increased neuronal cell death .
Cortical neural cells responded to NMDA receptor activation with an increase in mRNA for genes (BDNF, VGF, NARP, and IEGs such as Ania-3) known to be involved in synaptogenesis. We observed that the majority of these genes were upregulated 24 h after a brief (10 min) activation of NMDA receptors, indicating that a short event was enough to activate intracellular signal transduction pathways leading to gene transcription. Some of the IEGs and transcription factors (c-fos, NGFI-B) that we had previously reported as being significantly changed soon after 30–60 min treatment with NMDA , were still upregulated 24 h after a 10 min-treatment with NMDA. In addition to NGFI-B, NMDA treatment altered the expression of another member of the nuclear orphan receptor family, NOR-1, known to be activated by stress or neurotransmitters and to play a role in CNS function and perhaps in development [44–47]. NOR-1 null mice display developmental deficiency due to abnormal axonal guidance and impaired hippocampal development . NMDA treatment also increased the transcripts of the short inducible Homer family member, Ania-3, an IEG rapidly induced in response to different stimuli such as long-term potentiation, synaptic activity [49–52], traumatic brain injury  or drug of abuse (morphine, cocaine) [54–57]. Ania-3 expression is developmentally regulated by natural synaptic activity, and it plays a role in synaptogenesis [50, 51]. This IEG is involved in targeting of mGluRs to the plasma membrane, and binds to group 1 mGluRs . Homer proteins have been implicated in the pathology of neuropsychiatric disorders, such as epilepsy, schizophrenia, depression and addiction .
Collectively, our results show that treatment of young neuronal cultures with a high concentration of NMDA for short periods leads to plastic changes in neurons as a consequence of upregulation of IEGs and transcription factors. They will, in turn, induce changes in downstream genes known to participate in synapse organisation and formation. It is important to consider that since IEGs are still upregulated 24 h after NMDA-treatment, they might be activated in successive waves to trigger activation of “late-response” genes. Furthermore, since (i) IEGs are the most expressed genes during brain development [17, 60]; (ii) drugs of addictions, such as morphine, not only induce activation of transcription factors and IEGs, but also alter neurotransmitter receptor expression [54, 56, 61]; and (iii) all these changes could be long-lasting, then, they will certainly affect negatively a developing cerebral cortex. Hence, improper activation of IEGs or transcription factors may contribute to neurodevelopmental disorders following maternal exposure to neurotoxic drugs.
Activation of glutamatergic synapses in vivo induces rapid changes in gene expression subsequent to the activation of signaling cascades, which rapidly couples the signal of NMDA receptor activation at the synapse to the nucleus. Synaptic activity regulates gene expression through different pathways, and regulation of several genes is dependent on neuronal activity. Most of the activity-regulated genes encode for neurotrophic factors or transcription factors, which control expression of “late-response” genes [17, 21]. For example, BDNF mRNA is highly regulated by synaptic activity and the levels of neurotrophins are extremely important for cell survival during development .
Neurotrophin signalling is important during development for the formation and maturation of functional synapses. BDNF is rapidly upregulated by neuronal activity, and its levels are high during synaptic remodelling. Its involvement in LTP and in mediating synaptic strength and plasticity is well documented. Knockout mice for BDNF or its receptors TrkB/TrkC show reduced expression of synaptophysin, a protein associated with synaptic vesicles, reduced synaptic density, impaired LTP and deficits in presynaptic functions. These defects can be rescued by addition of exogenous BDNF. Conditional mutation in TrkB receptors have resulted in memory deficits [62–64].
Recently, VGF gene expression regulation has been linked to BDNF . Alder et al.  showed an upregulation of both VGF proteins and gene expression in hippocampal cultures treated with BDNF for 3 h. VGF is a polypeptide secreted during neuronal differentiation in cultures [65, 66], whose expression is regulated by neuronal activity, seizures and lesions . It was also reported that its mRNA colocalises with mRNA for TrkB receptors. Consequently, VGF may regulate release of other peptides  and may contribute to synaptic function by regulating release of BDNF and of other neuromodulators. Therefore, it is instructive to highlight the succession of the events observed in our study: activation of NMDA receptors for 180 min significantly upregulated BDNF (3.9-fold, Fig. 2A), this effect is followed by an increase in VGF expression (Fig. 2A). Both genes were still greatly expressed 24 h after a brief (10 min) activation of NMDA receptors (Fig. 3). Hence, we speculate that in cerebral cortex, BDNF may trigger changes in VGF gene expression, which in turn elicits synaptogenesis and synaptic function or simply potentiate BDNF effects. In line with this hypothesis, other genes that may play a role in process formation and remodeling, such as NF-L and vimentin (Fig. 3), were also upregulated. Furthermore, we found upregulation of NARP (Fig. 3), an IEG rapidly induced by synaptic activity that promotes neurite outgrowth and plays a role in excitatory synaptogenesis . Immunohistochemical studies have shown that NARP is expressed in hippocampus, cerebral cortex and habenula , and is selectively targeted to excitatory synapses, in tight association with the GluR1 subunit. Recently, it has been suggested that NARP participates in AMPA receptors clustering .
We also showed that NMDA receptor activation elicited a sustained increase in pCREB levels, suggesting a role for this transcription factor in mediating NMDA effects. The role of CREB in NMDA signalling is still controversial because of conflicting results reported by different groups . Several lines of evidence indicate that activation of NMDA receptors leads to CREB phosphorylation and activates CREB-dependent transcription. This idea is further supported by the fact that many of the genes activated following NMDA receptor stimulation contain a CREB binding site in their promoter region [40, 72]. Since some of the genes differentially expressed in our study (for example, c-fos, somatostatin, BDNF, Figs. 2, 3) are known to have a CREB response element, and we detected sustained levels of pCREB (Fig. 5), necessary to elicit activity-dependent transcription, long after the NMDA receptor activation had ceased, we propose that CREB is one of the central mediator of NMDA receptor effects in primary neuronal cultures. In addition, the high levels of pCREB could in part explain the lack of NMDA-induced apoptosis. CREB and related members have been implicated in control of neuronal adaptive responses, in regulation of BDNF transcription  and most importantly, in preventing neuronal death and in promoting survival in both developing and adult brain [73, 74].
Since, the synaptic vesicle protein synaptophysin is considered a marker of “synaptic density” and, since during brain development, its expression parallels the formation of synapses [42, 43], the increase in synaptophysin immunoreactivity displayed by neuronal cultures after treatment with NMDA further support our conclusions. Interestingly, we did not find a significant change in synaptophysin gene expression in our studies using the gene array Rat U34 neurobiology (data not shown). This result is in agreement with a previous report  showing that increases in the levels of synaptophysin are regulated at the translational level.
In our model, direct activation by NMDA of the neuronal program of maturation/development occurred at both transcriptional and post-transcriptional levels in the complete absence of astrocytes and oligodendrocytes, and therefore in the absence of glial-neuron interactions. However, by calcium imaging, we found that between 50% and 80% of the cells responded to NMDA with an increase in intracellular calcium (present paper and ). Thus, we could argue that the effects seen on gene expression, pCREB, cfos, and synaptophysin could be mediated by immediate and delayed responses of different types of progenitors and neuronal lineage cells present in the cultures to either glutamate or other neuromodulators (neuropeptides) released in the culture medium by the cells activated by NMDA.
In conclusion, we propose that (i) a massive activation of NMDA receptors due to high concentrations of NMDA may mimic the effects induced in vivo by repeated glutamate release following stimulation waves; (ii) like in the hippocampus, IEGs and transcription factors are key players in initiating cerebral cortical synaptic plasticity in response to NMDA; and (iii) our findings imply a key role for CREB in NMDA receptor actions, and they further support the contention for a connection between NMDA receptor, pCREB and plastic changes leading to proper synapses formation.
We thank Drs. David Krantz and Esteban Dell’Angelica for critically reading the manuscript. We would like to thank the NINDS/NIMH Microarray Consortium for assistance with expression profiling; and are also grateful to Donna Crandall and Carol Gray of the UCLA MRRC media core for their help with the figures. This work was supported by: the SNRP program at NINDS, NIH Grant NS-99002 (PAF and JDV); NIH grant HD-06576 (JDV); Staglin Music Festival and NARSAD Young Investigator Award, NIH Grant # DA-015802, Tobacco-Related Disease Research Program Grant # 11RT-0172, Alzheimer’s Association Grant # IIRG-023609 (DJS); NIH grant DA-05010 (AC).