| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
OPEN ACCESS ARTICLE
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TRANSLATIONAL AND CLINICAL RESEARCH |
aNeurodegeneration Department, Neuroscience Research, Novartis Institutes for BioMedical Research, Basel, Switzerland;
bBioMedical Computing, Genome & Proteome Sciences, Novartis Institutes for BioMedical Research, Basel, Switzerland;
cZoological Institute, TU Braunschweig, Braunschweig, Germany;
dIPMB, University of Heidelberg, Heidelberg, Germany;
eMPI for Brain Research, Frankfurt, Germany;
fBiocenter, University of Basel, Basel, Switzerland
Key Words. Amyloid precursor protein • Amyloid precursor-like protein 2 • Embryonic stem cells • Neuronal differentiation • Vesicular glutamate transporter 2 • Synaptic transmission
Correspondence: Miriam Bibel, Ph.D., Neurodegeneration Department, Neuroscience Research, Novartis Institutes for BioMedical Research, 4002 Basel, Switzerland. Telephone: +41-61-6966378; Fax: +41-61-6962809; e-mail: miriam.bibel{at}novartis.com
Received January 18, 2008;
accepted for publication May 8, 2008.
First published online in STEM CELLS EXPRESS June 5, 2008.
 |
ABSTRACT |
|---|
 TopFootnotes Abstract IntroductionMaterials and MethodsResultsDiscussionConclusionDisclosure of Potential...AcknowledgmentsReferences |
|---|
-secretase cleavage of APP in wild-type neurons resulted in a similar decrease of VGLUT2 expression, whereas VGLUT2 levels could be restored in App–/–Aplp2–/– neurons by a construct encompassing the C-terminal intracellular domain of APP. Electrophysiological recordings of hippocampal organotypic slice cultures prepared from corresponding mutant mice corroborated these observations. Gene expression profiling and pathway analysis of the differentiated App–/–Aplp2–/– neurons identified dysregulation of additional genes involved in synaptic transmission pathways. Our results indicate a significant functional role of APP and amyloid precursor-like protein 2 (APLP2) in the development of synaptic function by the regulation of glutamatergic neurotransmission. Differentiation of ESCs into homogeneous populations thus represents a new opportunity to explore gene function and to dissect signaling pathways in neurons. Disclosure of potential conflicts of interest is found at the end of this article.
|
INTRODUCTION |
|---|
|
TopFootnotesAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionDisclosure of Potential...AcknowledgmentsReferences |
|---|
We have recently established a system to differentiate ESCs into a homogeneous culture of glutamatergic neurons [1, 2]. Its homogeneity and synchrony allow us to analyze signaling pathways biochemically and to study key properties of neuronal development, function, and degeneration [3, 4].
Human genetics of Alzheimer's disease (AD) points to processing of the amyloid precursor protein (APP) as being directly implicated in AD pathology. However, the physiological function of APP and its proteolytic fragments is still largely unknown, in particular on the molecular level. APP is a member of a family of conserved type I transmembrane proteins including the amyloid precursor-like protein 1 (APLP1) and APLP2 in mammals. APP, APLP1, and APLP2 are highly homologous and similarly processed by several secretases. Cleavage of APP by 
- or β-secretase releases the soluble ectodomain of APP (APPs- and APPs-β) and generates C-terminal stubs (C-terminal fragment [CTF] and CTFβ). Subsequent intramembrane processing of CTFs by -secretase releases the APP intracellular domain (AICD) and amyloid β (Aβ), the major constituent of neuritic plaques in AD. In contrast, -secretase cleaves APP within the Aβ region, thereby preventing its formation. APP and APLP2 are expressed ubiquitously throughout the body in largely overlapping patterns during embryogenesis and in adult tissue [5], whereas APLP1 is found primarily in the nervous system, particularly in postsynaptic densities [6]. APP has been proposed to contribute to multiple cellular processes, including neurite outgrowth, neuronal survival, axonal transport, synaptogenesis, and transcriptional regulation [7–11]. Some of these functions have also been described for APLP1 and APLP2.
Mice deficient in one of the APP genes are viable and fertile. Only subtle neurological deficits are observed in mice deficient for App, and most of them can be rescued by re-expression of the APP N-terminal domains [12–14]. Combined gene mutations support partial functional complementation within the APP family and a key physiological role for APLP2 [15]. Indeed, while App–/–Aplp1–/– double mutants are apparently normal, App–/–Aplp2–/–, Aplp1–/–Aplp2–/–, and App–/–Aplp1–/–Aplp2+/– mutants die early after birth with no obvious histopathological abnormalities apart from defects at the neuromuscular junction [16].
Although β-amyloid plaques are a hallmark of AD, synaptic dysfunction may be an early and critical element in the pathogenesis of AD. APP and APLP2 show a strong expression increase at the time of synaptogenesis [17]. APP has also been suggested to participate in kinesin-mediated transport of vesicles, potentially including neurotransmitter vesicles [8]. Moreover, treatment of primary hippocampal cultures with APP antibodies leads to decreased spontaneous neuronal oscillations, probably resulting from fewer synapses being formed [18]. Recently, siRNA-targeted downregulation of APP and APLP2 in presynaptic terminals of rat retinal ganglion cells has been shown to result in a decrease of stimulus-driven synaptic activity [19]. Similarly, App–/– mice show deficits in synaptic function such as impaired long-term potentiation [20], and App–/–Aplp2–/– mice present alterations in synaptic structure and transmission at neuromuscular junctions and submandibular ganglia [16, 21]. Therefore, both proteins may have a role in synapse formation and function.
In this study, we compared the expression pattern of synaptic proteins in wild-type and App–/–Aplp2–/– neurons using our system of synchronous neuronal differentiation of mouse ESCs [1]. The homogeneity of this system allowed us to identify a molecular link between APP and APLP2 proteins and synaptic transmission by showing that expression of the vesicular glutamate transporter 2 (VGLUT2) protein is downregulated in the absence of both APP and APLP2. Moreover, glutamate release and synaptic transmission are impaired when these proteins are not present.
|
MATERIALS AND METHODS |
|---|
|
TopFootnotesAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionDisclosure of Potential...AcknowledgmentsReferences |
|---|
Differentiation protocol was as published in Bibel et al. [1, 2]. Neurons were cultured in a medium based on the one described by Brewer and Cotman [23], resulting in a homogeneous culture of glutamatergic neurons.
Immunocytochemistry
Glass coverslips were acid treated before usage, and cells were fixed with 4% paraformaldehyde for 10 minutes. Blocking was for 1 hour in blocking buffer (0.03% carrageenan, 10% normal goat serum (NGS), 0.3% Triton X-100 (Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com) in phosphate-buffered saline [PBS]), and incubation was with the respective antibodies at 4°C overnight in incubation buffer (0.03% carrageenan, 3% NGS, 0.3% Triton X-100 in PBS). Mounting was in AquaPoly/Mount (Polysciences Inc., Warrington, PA, http://www.polysciences.com). For immunocytochemistry the following antibodies were used: nestin (rat401, monoclonal IgG1; 1:10), RC2 (monoclonal IgM; 1:4), Pax6 (monoclonal IgG1; 1:100; all purchased from Developmental Studies Hybridoma Bank at University of Iowa, Iowa City, IA http://dshb.biology.uiowa.edu), microtubule-associated protein 2 (MAP2) (monoclonal IgG1, clone HM-2; 1:250; Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com), and Neurofilament M (rabbit polyclonal; 1:200; Chemicon, Temecula, CA, http://www.chemicon.com). Secondary antibodies conjugated with cyanin 2 (Cy2) or Cy3 were used (1:500; Jackson Immunoresearch Laboratories, West Grove, PA, http://www.jacksonimmuno.com).
Western Blot Analysis
Protein lysates from cells were prepared using lysis buffer (50 mM Trizma base, 150 mM NaCl, 10% glycerol, 1% Triton X-100) supplemented with protease inhibitor cocktail (Roche, Basel, Switzerland, http://www.roche-applied-science.com). Western blots were performed with the NuPAGE System and the XCell II Blot Module, both from Invitrogen (Carlsbad, CA, http://www.invitrogen.com) on polyvinylidene fluoride membranes. Blots were blocked with a 5% milk solution (Tris-buffered saline Tween-20, 0.2% Tween, [Sigma Aldrich]); detection was performed with ECL Plus (Amersham Biosciences, Piscataway, NJ, http://www.amersham.com). The following primary antibodies were used: rabbit polyclonal antibodies to APLP1 (1:3000; Calbiochem, San Diego, http://www.emdbiosciences.com), APLP2 (1:5000; Calbiochem), APP C8 (1:3000; raised against the 19 C-terminal amino acids of human APP that cross-reacts with mouse APP; kindly provided by P. Paganetti, Novartis, Basel, Switzerland), VGLUT2 (1:5000; Synaptic Systems, Goettingen, Germany, http://www.sysy.com), glutamate receptor subunit 1 (GluR1, 1:1000; Upstate, Charlottesville, VA, http://www.upstate.com), and the monoclonal antibodies to synaptophysin (1:2000; Sigma-Aldrich), synaptobrevin/VAMP2 (1:10000; Synaptic Systems), postsynaptic density protein 95 (PSD-95, 1:1000; Sigma-Aldrich), tau (
-1, 1:3000; Chemicon), and β-tubulin isotype III (1:5000; Sigma-Aldrich).
Reverse-Transcription–Polymerase Chain Reaction
Total RNA was extracted and purified according to the protocol of Chomczynski and Sacchi [24], using TRIZOL (Invitrogen). RNA was quantified using the ND-1000 machine (NanoDrop Technologies; Thermo Fisher Scientific Inc., Waltham, MA, http://www.thermofisher.com). For cDNA synthesis 400 ng of DNase-treated total RNA was used with 300 ng of random primers. The polymerase chain reaction (PCR) reactions were performed at 95°C for 1 minute, 63°C for 1 minute, 72°C for 1 minute during 34 cycles followed by a final elongation step of 2 minutes at 72°C. Products were analyzed on a 1.5% agarose gel stained with ethidium bromide.
Quantitative Reverse-Transcription-PCR
Quantitative reverse-transcription (RT)-PCR was performed using TaqMan probes (Applied Biosystems, Foster City, CA, http://www.appliedbiosystems.com).
From each RNA sample, two to three independent cDNAs were synthesized and two to three independent quantitative PCRs (qPCRs) were performed with each cDNA in triplicates. Five ng cDNA was taken for 20 µl reaction volume. qPCR was performed in 96-well format with an ABI Prism SDS 7900 HT thermocycler (Applied Biosystems). Controls without RT were included to check for traces of genomic DNA. The amplification protocol was as follows: 50°C for 2 minutes, 95°C for 10 minutes, and 40 cycles at 95°C for 15 seconds and at 60°C for 1 minute. Quantification was done according to Pfaffl [25].
Construction of APP ICD Targeting Vector
A targeting vector for the Mapt locus was constructed containing pgk-puro-pA that placed the cDNA into exon1 of the tau gene. The intracellular domain of human APP (huAPP ICD) was cloned (amino acids [aa's] 680–751) in NheI/NotI of the pRC/CMV AC7 vector, a derivative of the pRC/CMV vector (Invitrogen), which contains the BM40 signal peptide [26]. BM40 huAPP ICD was then subcloned into PmeI/NotI of tau targeting vector with an optimized Kozak sequence.
Gene Targeting in Mouse ESCs
The targeting vector was linearized by digestion with AscI. App–/–Aplp2–/– ESCs were electroporated with 30 µg linearized DNA using a Bio-Rad Gene Pulser apparatus (Hercules, CA, http://www.bio-rad.com). Electroporated cells were plated on MEFs and selected for 1 week with puromycin (1 µg/ml). One hundred twenty clones were picked and three clones were detected positive by PCR and Southern blotting for insertion of huAPP ICD into the Mapt locus.
Electrophysiological Recordings
Hippocampal organotypic slice cultures were prepared from mice at embryonal day 19 (E19) or postnatal day 1 (P1) following the method of Stoppini et al [27]. After 14–20 days in vitro (DIV), slices were transferred to a recording chamber and kept submerged at 32°C while constantly being perfused (1 ml/min) with aerated artificial cerebral spinal fluid. Schaffer collaterals in area cornus ammonis 3 (CA3) were stimulated by monopolar tungsten electrodes, and field excitatory postsynaptic potentials (EPSPs) were recorded by a glass electrode placed in the stratum radiatum of CA1. Data were collected with a LabView (National Instruments, Munich, Germany, http://www.ni.com) based acquisition program, and amplitudes of the field EPSPs (fEPSPs) were calculated. Input/output curves were collected with a stimulus strength ranging from 20–80 µA to adjust the fiber volley size between 0.1–0.8 mV.
Glutamate Release Assay
Glutamate release assay was performed as described in Griesbeck et al. [28], with slight modifications. In brief, neurons derived from wild-type and App–/–Aplp2–/– ESCs were cultured on 12-well plates for 14 days (14 DIV). They were incubated with 1 µl 3[H]-glutamate (L-[G-3H] glutamic acid, 9.25 MBq; Amersham Biosciences) for 1 hour at 37°C. After washing cells twice with prewarmed Hank's buffer, medium was collected every 5 minutes and replaced by fresh medium. After 25 minutes (after fraction no. 5) and 45 minutes (after fraction no. 9), 50 mM KCl or 100 µM glutamate was added. Medium was collected up to fraction no. 12 and the amount of 3[H] was determined by a scintillation counter. For data analysis the amount of glutamate being released was quantified as follows: c – [(a + b)/2], where a indicates counts of fractions nos. 4 and 5 before stimulation; b, counts of fractions nos. 7 and 8 after stimulation; and c, of fraction no. 6 directly after stimulation.
|
RESULTS |
|---|
|
TopFootnotesAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionDisclosure of Potential...AcknowledgmentsReferences |
|---|
ESCs derived from double mutant and wild type differentiated in a similar fashion according to the protocol in Bibel et al. [1, 2]. As APP is strongly expressed by radial glial cells, the neuronal progenitors, we first examined the ability of ESCs lacking both genes to differentiate into radial glial cells by immunocytochemistry using nestin, RC2, and Pax6 antibodies. Neural progenitors of double-mutant ESCs compared with wild-type ESCs displayed a similar morphology (Fig. 1A), and the percentage of nestin-, RC2-, and Pax6-positive cells after 2–5 and 24 hours was not different between the double mutants and wild type (Fig. 1B). Also, differentiation of the neural progenitors into mature neurons was not altered in the absence of APP and APLP2; neither viability in culture nor neurite outgrowth and branching were affected, as detected by immunocytochemistry for MAP2 and neurofilament (Fig. 1C). No TH, Isl1, or Chat RNA could be detected after 12 days in culture, and VGAT RNA expression resulting from approximately 5% of GABAergic neurons [1] was not altered (supplemental online Fig. 1B); all neurons, however, expressed VGLUT, corresponding to the characterization described in Bibel et al. [1]. However, VGLUT expression was determined to correspond to VGLUT2 and not to VGLUT1 as previously assumed. Indeed, we could not detect VGLUT1 with qPCR (supplemental online Fig. 1A) or by Western blots with VGLUT1-specific antibodies. Therefore, the specific immunostaining for VGLUT1 reported in our previous publications [1, 2, 4] must be the result of a cross-reactivity of the VGLUT1 antibody. In line with this, the antigen used to raise this antibody to VGLUT1 shows more than 80% identity with VGLUT2. These results show that the deletion of App and Aplp2 does not affect the generation of Pax6-positive radial glial cells or their subsequent differentiation into glutamatergic neurons.
|
App–/–Aplp2–/– Neurons Display a Strong Deficit in VGLUT2 Expression
A key feature of our differentiation protocol is the synchronous differentiation of neurons from their progenitors, thus allowing a precise monitoring of the levels of synaptic components including synaptophysin (Syp), synaptobrevin, VGLUT2, PSD-95, and GluR1. Comparisons between wild-type and mutant cells were made at 12 days in vitro by Western blot, as previous electrophysiological recordings have established that functional synapses can be detected at this time. To account for possible differences in neuronal density, protein amounts were adjusted to equal levels of βIII-tubulin expression. Whereas similar levels of synaptophysin, Mapt, and βIII-tubulin expression were observed in cultures of either genotype, a remarkable reduction of VGLUT2 expression was observed in neurons lacking App and Aplp2 (Fig. 2A). Expression of both synaptobrevin and GluR1 was also similar between the genotypes, whereas that of PSD-95 was also reduced in App–/–Aplp2–/– neurons (Fig. 2C).
|
Inhibition of -Secretase Cleavage of APP Proteins Reduces VGLUT2 Expression
The reduction of VGLUT2 transcript levels in cells lacking App and Aplp2 suggests that the amyloid precursor proteins may be directly involved in the regulation of transcription of genes such as VGLUT2. The intramembrane cleavage of the APP proteins by -secretase leads to the release of small intracellular fragments (AICD and APLP intracellular domain [ALICD]), which have been previously shown to enter the nucleus and to participate in the formation of a transcriptional regulatory complex [10, 32]. To test whether these peptide fragments play a role in VGLUT2 regulation, we blocked the generation of the intracellular fragments by treating wild-type neurons with a highly potent -secretase inhibitor, LY411575 (50 nM), which has a median inhibitory concentration <1 nM in human embryonic kidney cells [33]. Enzyme-linked immunosorbent assay measurements of Aβ40 levels indicated that the inhibition of -secretase was complete (data not shown). -Secretase is known to cleave other transmembrane proteins such as Notch, which is involved in cell-fate determination. Therefore, we treated our neurons with the -secretase inhibitor after 6 days of differentiation so as to minimize any possible impairment of Notch function during the generation of neurons. Cultures were treated either on day 6, or on day 6 and day 11, and RNA was extracted at day 12 and the levels of VGLUT2 and Mapt were measured by qPCR. The morphology of neurons treated with the -secretase inhibitor was indistinguishable from that of either wild-type or double knockout neurons (Fig. 1C). We also tested cell fate and determined that it was not changed as assessed by Th, Isl1, Chat, and VGAT RNA expression (supplemental online Fig. 1B). -Secretase inhibition caused a highly significant decrease in VGLUT2 expression (Fig. 3B), whereas Mapt mRNA levels were not affected (Fig. 3A). Thus, the cleavage of APP proteins by a -secretase-dependent mechanism is necessary for VGLUT2 transcription to proceed normally.
|
|
|
In wild-type neurons, a high level of 3[H]-glu release was detected after the first stimulation with 100 µM glutamate (Fig. 5B), although 3[H]-glu release after the second stimulation was much lower due to depletion. In App–/–Aplp2–/– neurons, 3[H]-glu release was clearly reduced after both stimulations with 100 µM glutamate (Fig. 5C). Release of 3[H]-glu in App–/–Aplp2–/– neurons after the first stimulation with either 50 mM KCl (Fig. 5D) or 100 µM glutamate (Fig. 5E) was approximately 50% of that observed in wild-type neurons. In both genotypes, a slow decline of 3[H]-glu was observed over time; this is most likely due to synaptic release in response to spontaneous neuronal activity. These results indicate that glutamate uptake and release are decreased in double knockout neurons.
Gene Expression Profiling Indicates Defects in Synaptic Transmission Pathways in App–/–Aplp2–/– Neurons
To explore whether the expression levels of additional genes involved in synaptic transmission were also affected in App–/–Aplp2–/– neurons, two independent neuronal differentiations of wild-type and App–/–Aplp2–/– ESCs were performed, RNA was extracted after 12 days and subjected to DNA microarray analysis (supplemental online Table 1). Besides detecting the lack of App and Aplp2, the mouse gene expression profiling confirmed the downregulation of VGLUT2 (Slc17a6), and identified other genes involved in synaptic transmission, such as synaptic vesicle glycoprotein 2c (Sv2c) and neuritin 1 (Nrn1), a gene induced by neural activity (Fig. 6A). Also of particular interest was visinin-like protein 1 (Vsnl1), previously described to be associated with amyloid plaques and tangles in AD [31]. In addition, several strongly regulated genes of unknown function were identified, such as Btbd9, which contains a DNA-binding domain. These transcriptional changes were also confirmed by qPCR (Fig. 6B).
|
|
|
DISCUSSION |
|---|
|
TopFootnotesAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionDisclosure of Potential...AcknowledgmentsReferences |
|---|
-secretase inhibition in wild-type neurons considerably reduces the expression of VGLUT2. The prediction that the decreased vesicular uptake of glutamate in ESC-derived neurons alters neurotransmission in vivo is supported by the electrophysiological recordings performed in hippocampal organotypic slice cultures of double-mutant animals. APP/APLP2-controlled expression of VGLUT2 may play an important role in synaptic plasticity and its dysregulation in AD. A prominent role of the APP family in neuronal and synaptic function is further underlined by gene expression profiling analysis that identifies genes involved in neuronal activity and synaptic transmission and reveals down-regulation in neurons lacking App and Aplp2.
APP Proteins Regulate VGLUT2 Expression and Affect Synaptic Transmission
VGLUT2 is one of three recently identified vesicular glutamate transporters [35, 36]. VGLUT1 and VGLUT2 show a complementary expression pattern in the adult brain defining distinct subsets of excitatory glutamatergic neurons [37], whereas VGLUT3 is expressed by many cells releasing transmitters other than glutamate [38]. In rodents, VGLUT2 is already expressed at early developmental stages, whereas in some brain regions VGLUT1 transcription increases postnatally while VGLUT2 expression decreases, suggesting a switch of VGLUT isoforms [39]. At the maturation stage of the ES-derived neurons used in this study, a possible switch of VGLUT2 to VGLUT1 expression has not taken place, and therefore it is unknown if VGLUT1 expression in postnatal neurons is similarly regulated. Interestingly, the targeted deletion of VGLUT1 in mice causes postnatal lethality [40], possibly indicating that VGLUT2 can only partially compensate for the lack of VGLUT1. Also deletion of VGLUT2 causes postnatal lethality in accordance with the importance of VGLUT2 in early development, and defects in the heterozygous animals are in line with the localization of VGLUT2 [41]. Since the amount of glutamate transported in synaptic vesicles depends on the number of VGLUT molecules present per vesicle [41, 42], expression levels of VGLUTs control presynaptic regulation of quantal size and the efficacy of neurotransmission [42–44]. A single transporter unit may be sufficient to fill a synaptic vesicle [45], thus suggesting that the precise regulation of VGLUT expression is of critical physiological relevance.
Consistent with this idea, we find impaired neurotransmission by performing electrophysiological recordings on organotypic hippocampal slice cultures of App–/–Aplp2–/– animals. This impairment is likely due to reduced presynaptic release as a result of reduced levels of VGLUT2 expression as App–/–Aplp2–/– ESC-derived neurons showed a reduced release of glutamate compared with controls in response to two independent methods of stimulation, either with glutamate or KCl, and no change in GluR1 protein expression was detected in ESC-derived neurons. However, at this stage it cannot be excluded that reduced glutamate uptake results in decreased synapse numbers or glutamate receptor trafficking, which would also lead to smaller fEPSPs. The data suggest, however, that the largest defect is presynaptic in nature.
The APP Intracellular Domain Is Crucial for Transcriptional Regulation
Our findings suggest a requirement of AICD in the regulation of VGLUT2 transcription since preventing the cleavage of APP by inhibition of -secretase activity in wild-type cells mimics the results obtained with the App–/–Aplp2–/– cells. Furthermore, we demonstrate that expression of the AICD via a CT71 construct in App–/–Aplp2–/– neurons is sufficient to restore transcription of VGLUT2, although the mode of its interaction with Fe65 in this context, if any, remains unclear. As the short time period during which AICD levels can be detected coincides with the onset of synapse formation [46], our results suggest that this domain plays an important role for the appropriate developmental expression of VGLUT2.
AICD and ALICD can be stabilized by members of the Fe65 family, which are cytoplasmic scaffolding proteins that are colocalized with growth-associated protein 43 in growth cones and synapses, and are associated with rab-5 vesicles [47]. Fe65 is essential for proper cortical development, as the targeted deletion of two members of this family, Fe65 and Fe65L, results in type II lissencephaly [48], a cortical dysplasia reminiscent of that observed in App triple-mutant mice [49]. Fe65 is presumably activated by binding to APP, and then translocates to the nucleus to form a transcriptionally active complex with Tip60 [10, 11, 50].
Our gene profiling data analysis of App–/–Aplp2–/– neurons identifies a coordinated down-regulation of genes involved in neuronal activity and synaptic transmission. Previous studies have identified several candidate AICD target genes, including CD82 (Kai1), glycogen synthase kinase-3beta (Gsk3b), Neprilysin, Rab3b, and LRP1 [51–56]. However, other studies could not confirm the regulation of these genes [57]. Although our results confirm down-regulation of Gsk3b, Rab3b, and other genes involved in synaptic transmission in the App–/–Aplp2–/– neurons, the transcription of several of the previously reported target genes is either not altered or even sometimes increased. These discrepancies are most likely explained by different experimental systems and approaches used by the different investigators. In contrast to other studies where precisely timed inhibition of -secretase was used to identify acutely regulated genes, App and Aplp2 and their cleavage products were absent during the entire development of the neurons in our experimental paradigm. Other studies used a high overexpression of a non-membrane-bound form of AICD to look for large-scale gene regulation. It has been shown that the intracellular domain of APP, before being cleaved to release the AICD, is able to bind Fe65 and recruit it to the cytoplasmic membrane [32, 47, 58], and that the phosphorylation of APP at Thr668 is essential for the regulation of Fe65 nuclear translocation and activity [59]. Consequently, it is possible that overexpression of a non-membrane-bound form of AICD could disturb the normal phosphorylation of APP or block the signaling of the APP proteins by dimerizing with their intracellular domains, binding Fe65 in a way that would prevent its translocation to the nucleus, thus possibly resulting in the same gene expression alterations as those induced by the lack of App and Aplp2.
A key difference with previous work is the fact that our data were obtained from a homogenous population of glutamatergic neurons displaying the characteristics of excitatory neurons including spontaneous and inducible activity, whereas all other studies were performed with either cell lines or whole brain lysates.
The role of the AICD in the physiological function of the APP family might not be cell autonomous, but may be regulated by the extracellular domains to which it is attached as shown in Caenorhabditis elegans by Hornsten et al. [60] or by interaction partners that directly mediate transcriptional regulation.
Function of the APP Gene Family in Neuronal Development and Disease
Mice deficient for individual or all possible combinations of App family members indicate functional redundancy between Aplp2 and its isoforms, suggesting a key role of Aplp2. None of the double mutants that died early showed any obvious histopathological abnormalities in the brain or any other organ examined, whereas App–/–Aplp1–/–Aplp2–/– mice die shortly after birth with 81% showing cranial abnormalities, mainly cortical dysplasia resembling human type II lissencephaly and partial loss of Cajal Retzius cells, supporting an essential role of the App family in neuronal development [49].
APP and APLP2 are expressed in largely overlapping patterns, both at cellular and subcellular levels. Colocalization has been shown in CA1 pyramidal neurons [5], which share some characteristics with our ESC-derived neurons, such as the use of glutamate as neurotransmitter. Although our cells also express Aplp1, its expression level is not influenced by deletion of App or Aplp2, suggesting that even if the ALICD of APLP1 participates in VGLUT2 transcription, it may not suffice on its own to achieve normal expression levels.
In our cultures, we did not observe any alterations in the characteristics of differentiation or the survival rate of neurons derived from mutant ESCs, consistent with previous reports of the cortex of App–/–Aplp2–/– mice at the light and electron microscopic level [15]. We note that recent studies with App–/– Aplp2–/– mice at birth indicate reductions in the density of synaptic vesicles and the size of active zones in the submandibular ganglion [21], and at the neuromuscular junction, where neurotransmitter release has also been found defective [16]. Although the detailed mechanisms remain unclear, our results and these other studies clearly indicate an important role of the APP protein family in synaptic function.
|
CONCLUSION |
|---|
|
TopFootnotesAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionDisclosure of Potential...AcknowledgmentsReferences |
|---|
|
DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST |
|---|
|
TopFootnotesAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionDisclosure of Potential...AcknowledgmentsReferences |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TopFootnotesAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionDisclosure of Potential...AcknowledgmentsReferences |
|---|
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TopFootnotesAbstractIntroductionMaterials and MethodsResultsDiscussionConclusionDisclosure of Potential...AcknowledgmentsReferences |
|---|
is sufficient to rescue the anatomical, behavioural, and electrophysiological abnormalities of APP-deficient mice. J Neurosci 2007;27:7817–7826.
50 kDa; APP, between 100–120 kDa; APLP1, at 95 kDa; and APLP2, at 110 kDa. APLP1 showed no upregulation in dko neurons compared with wild-type neurons. (E): Reverse-transcription–polymerase chain reaction of App, Aplp1, and Aplp2. RNA was isolated from wild-type and two independent differentiations of dko neurons after 12 DIV. Primers used for App detection amplify a fragment of 450-bp length, primers for Aplp1 a fragment of 750-bp length, and primers for Aplp2 detection a fragment of 464-bp length. Mapt was used as housekeeping gene and primers for its detection amplify a fragment of 422-bp length. Note that the levels of Aplp1 appear to be similar in both genotypes.
