A Critical Role for System A Amino Acid Transport in the Regulation of Dendritic Development by Brain-derived Neurotrophic Factor (BDNF)*

Dendritic development is essential for the establishment of a functional nervous system. Among factors that control dendritic development, brain-derived neurotrophic factor (BDNF) has been shown to regulate dendritic length and complexity of cortical neurons. However, the cellular and molecular mechanisms that underlie these effects remain poorly understood. In this study, we examined the role of amino acid transport in mediating the effects of BDNF on dendritic development. We show that BDNF increases System A amino acid transport in cortical neurons by selective up-regulation of the sodium-coupled neutral amino acid transporter (SNAT)1. Up-regulation of SNAT1 expression and System A activity is required for the effects of BDNF on dendritic growth and branching of cortical neurons. Further analysis revealed that induction of SNAT1 expression and System A activity by BDNF is necessary in particular to enhance synthesis of tissue-type plasminogen activator, a protein that we demonstrate to be essential for the effects of BDNF on cortical dendritic morphology. Together, these data reveal that stimulation of neuronal differentiation by BDNF requires the up-regulation of SNAT1 expression and System A amino acid transport to meet the increased metabolic demand associated with the enhancement of dendritic growth and branching.

differentiation of specific neuronal populations in the peripheral and central nervous system (2). In the developing visual cortex, BDNF has been shown to regulate the dendritic growth and complexity of pyramidal neurons (3,4). Furthermore, overexpression of BDNF in pyramidal neurons of the visual cortex leads to sprouting of basal dendrites and regression of dendritic spines (5). Because dendritic morphology determines the number, pattern, and types of synapses received by a neuron, regulation of cortical dendritic growth and branching by BDNF is likely to play a major role for the proper functioning of the brain and especially the cerebral cortex.
Although BDNF regulates cortical dendritic development, little is known about the cellular and molecular mechanisms underlying these effects. In particular, the role of amino acid transport in mediating the effects of BDNF on dendritic growth and branching remains unknown.
System A is a ubiquitous amino acid transport system that mediates the Na ϩ -dependent transport of short-chain neutral amino acids such as alanine, serine, and glutamine (6). In addition, System A is the major amino acid transport system subject to regulation by environmental conditions, proliferative stimuli, developmental changes, hormones, and growth factors (6). During gestation, inhibition of System A-mediated amino acid transport is associated with intrauterine growth retardation (7). Recently, three isoforms of the System A family of transporters have been cloned and termed sodium-coupled neutral amino acid transporter (SNAT)1, SNAT2, and SNAT4 (8). SNAT1 is found primarily in the brain, SNAT2 is expressed ubiquitously in mammalian tissues, and SNAT4 is restricted almost exclusively to the liver (8). A role has been proposed for SNAT1 in supplying glutamatergic and GABAergic neurons with glutamine, the preferred precursor for the synthesis of the neurotransmitters glutamate and GABA (9,10). However, recent data have revealed that SNAT1 is not observed in axon terminals but is localized almost exclusively to the somatodendritic compartment of glutamatergic and GABAergic neurons (11), indicating that the physiological role of SNAT1 in the brain remains controversial.
In the present study, we examined the role of System A and SNAT1 in the regulation of cortical dendritic development by BDNF. We show that stimulation of cortical neurons by BDNF up-regulates SNAT1 expression and System A amino acid transport. Induction of SNAT1 expression and System A activity is required for the increase in dendritic length and branching * This work was supported by Swiss National Science Foundation Grant 3200BO-103652 (to J.-L. M.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 To whom correspondence should be addressed. of cortical neurons in response to BDNF. We also provide evidence that up-regulation of SNAT1 expression and System A activity by BDNF is necessary to enhance levels of tissue-type plasminogen activator (tPA), a protease whose secretion is essential for the effects of BDNF on dendritic morphology.

EXPERIMENTAL PROCEDURES
Cell Culture-All experiments were performed in accordance with the European Communities Council Directive regarding the care and use of animals for experimental procedures. Primary cultures of cerebral cortical neurons were prepared from 17-day-old Swiss mice embryos as described previously (12). If not otherwise indicated, cells were plated at a density of 55,000/cm 2 on culture plates or glass coverslips in Neurobasal medium supplemented with B27 (Invitrogen, Basel, Switzerland) and 500 M glutamine. Human embryonic kidney cells (HEK293; Invitrogen) were cultured in Dulbecco's modified Eagle's medium (Invitrogen) containing 10% fetal bovine serum (Invitrogen), 25 mM glucose, 2 mM glutamine, 0.1 mM nonessential amino acids (Invitrogen), 50 units/ml penicillin, 50 mg/ml streptomycin and kept at 37°C in a humidified atmosphere at 95% air, 5% CO 2 .
Dendritic Analysis-Cortical neurons were grown at a density of 1400 cells/cm 2 to allow morphological analysis of single neurons. After 3 days in culture, cortical neurons were exposed to 10 ng/ml BDNF with or without 20 mM MeAIB or 2 M tPA stop and immunostained for MAP2. Neurons were traced in a blind manner using a microscope equipped with ϫ63 and ϫ100 objectives, epifluorescence, a Lucivid camera system (Micro-BrightField Inc., Williston, VT), and the Neurolucida software (MicroBrightField Inc.). As Sholl analysis of preliminary studies revealed that the effects of BDNF on dendritic length and branching were maximal within a radius of 80 m from the cell body (not shown), data presented in this study include analysis of dendritic morphology within this radius. When the effects of SNAT1 siRNA and control (Ctrl) siRNA were tested, cortical neurons were cotransfected 2 days after plating with the enhanced green fluorescent protein (EGFP) expression vector pEGFP (0.1 g/well; Clontech, Basel, Switzerland) together with SNAT1 siRNA or Ctrl siRNA (0.5 g/well). One day after transfection, neurons were treated for 24 h with 10 ng/ml BDNF in Neurobasal/B27 medium and fixed with 4% paraformaldehyde. EGFP-positive neurons were traced and analyzed using Neurolucida software.
Small Interfering (si) RNAs-SNAT1 and control (no significant sequence homology to any known mammalian gene) 21-nucleotide duplexes were designed and synthesized by Qiagen (Basel, Switzerland). The target sequence for SNAT1 siRNA was 5Ј-AACGAACTTCCTTCAGCCATA-3Ј, corresponding to nucleotides 511-531 of the mouse SNAT1 cDNA sequence.
Transfection-Transfection of plasmid as well as cotransfection of plasmid with siRNA were performed using the Transfectin reagent (1 l/well in 24 well plates; Bio-Rad, Glattbrugg, Switzerland) according to the manufacturer's guidelines. When siRNA duplexes (2.5 g/35 ϫ 10-mm plate) were transfected alone, GeneSilencer reagent (Gene Therapy Systems, Axon Lab, Baden-Dättwil, Switzerland) was used following the manufacturer's manual.
LDH Release-LDH release into the culture supernatant was quantified using a sensitive detection kit (Cytotoxicity Detection kit Plus (LDH), Roche Diagnostics, Mannheim, Germany).
MTT Reduction-Measurement of cellular MTT reduction was performed as described (14).
Protein Synthesis-Analysis of protein synthesis was performed as described previously (13).
Zymography-tPA activity was assayed by zymography as described previously (12). Bands of caseinolysis at ϳ68 kDa corresponding to the molecular weight of tPA were analyzed using ScionImage Software, version Beta 4.0.2 (Scion Corp., Frederick, MD).
Northern Blotting-Northern blots were carried out using a previously described procedure (12). 32 P antisense SNAT1 and SNAT2 riboprobes were generated by in vitro transcription of a linearized pGEM-T Easy vector containing the corresponding cDNA fragment (see "RT-PCR and DNA Constructs").
Statistical Analysis-Data were analyzed for statistical significance by using one-way analysis of variance followed by Bonferroni post hoc test except for time course analysis of MeAIB uptake where one-way analysis of variance was followed by Dunnett post hoc test.

Regulation of System A Amino Acid Transport by BDNF-
System A can be distinguished from the other neutral amino acid transport systems by its ability to transport N-methylated amino acids such as the nonmetabolizable amino acid analog MeAIB (15). Regulation of System A amino acid transport activity by BDNF was examined by measuring MeAIB uptake. We found that treatment of cortical neurons with BDNF caused a marked increase in MeAIB uptake (Fig. 1). Time course analysis revealed that MeAIB uptake enhanced significantly after 20 min and reached maximal levels (179.4 Ϯ 7.6% of control) after 5 h of stimulation by BDNF (Fig. 1A). These data indicate that BDNF increases System A activity in cortical neurons. In neurons, System A plays an important role in the Na ϩ -dependent transport of glutamine, the primary precursor for the synthesis of the neurotransmitters glutamate and GABA (16). This led us to examine the effect of BDNF on glutamine transport, and we found that BDNF also markedly increased glutamine uptake by cortical neurons (Fig. 1B). Regulation of System A activity by BDNF was further characterized by performing kinetic analysis of MeAIB uptake in control and BDNF-treated cortical neurons. The rates of MeAIB uptake in control and BDNF-treated cultures exhibited Michaelis-Menten kinetics (Fig. 1C). Double-reciprocal plots of MeAIB uptake rate versus extracellular MeAIB concentration revealed that the V max of MeAIB uptake increased significantly from 10.5 Ϯ 0.9 to 18.3 Ϯ 2.5 pmol/mg of protein/min when cortical neurons were exposed to BDNF (Fig. 1D). In contrast, no significant effect was observed on the K m value as K m was 447.4 Ϯ 14.2 M in control neurons and 410.5 Ϯ 65.6 M in BDNF-treated neurons (Fig. 1D). The increased V max value of MeAIB uptake in response to BDNF provides evidence that BDNF enhances the number of functional System A transporters in cortical neurons. Because we have previously shown that treatment of cortical neurons with BDNF increases overall protein synthesis (13), we examined the role of System A amino acid transport in BDNF-induced protein synthesis. Treatment of cortical neurons with an excess of MeAIB to competitively inhibit System A amino acid transport caused a complete suppression of BDNF-induced protein synthesis (Ctrl ϭ 100 Ϯ 2%, BDNF ϭ 129.8 Ϯ 2.1%, BDNF ϩ MeAIB ϭ 92.45 Ϯ 1.99%, n ϭ 12). These data indicate that up-regulation of System A amino acid transport is necessary for the increased protein synthesis by BDNF.
Up-regulation of System A Amino Acid Transport Is Required for the Regulation of Dendritic Morphology by BDNF-Because BDNF regulates dendritic development of cortical neurons (3,4,17), we investigated whether up-regulation of System A transport contributes to the effects of BDNF on the dendritic morphology of cortical neurons. To address this issue, cortical neurons were stimulated by BDNF in the presence of the amino acid analog MeAIB, and the dendritic morphology was analyzed. When compared with untreated cultures, neurons exposed to BDNF exhibited an increased number of branch points (Fig. 2, A and B). In marked contrast, treatment of cortical neurons with MeAIB completely suppressed the effect of BDNF on dendritic branching (Fig. 2, A and B), indicating that stimulation of System A amino acid transport is required for the regulation of cortical dendritic morphology by BDNF. Suppression of BDNF-induced dendritic development by MeAIB did not result from a decrease in neuronal viability, as determined by LDH release and MTT reduction (Fig. 3, A and B). Further analysis of the regulation of dendritic branching by BDNF showed that BDNF increased the number of branch points in both GABAergic (21.5% of the total neuronal population) and non-GABAergic neurons, although the effect of BDNF was more pronounced in non-GABAergic cells (Table 1). In addition, inhibition of System A amino acid transport by MeAIB suppressed BDNF-induced dendritic branching in both populations of cortical neurons (Table 1).
BDNF Increases SNAT1 Expression-To investigate further the role of System A in the regulation of neuronal differentiation by BDNF, we examined the System A amino acid transporter isoform involved in the effects of BDNF on the dendritic development of cortical neurons. In agreement with the tissue distribution of SNAT isoforms (8), RT-PCR analysis revealed that SNAT1 and SNAT2 mRNAs were expressed in cultured cortical neurons, whereas SNAT4 mRNA was not detected (Fig.  4A). On the basis of these findings, we tested the effect of BDNF on the expression of SNAT1 and SNAT2. Treatment of cortical neurons with BDNF caused an up-regulation of SNAT1 mRNA expression, whereas SNAT2 mRNA levels remained unchanged (Fig. 4, B and C). Consistent with increases in SNAT1 mRNA levels, we found that BDNF enhanced SNAT1

Role of System A amino acid transport in BDNF-induced dendritic branching of cortical GABAergic and non-GABAergic neurons
Cortical neurons were treated with 10 ng/ml BDNF for 24 h in the presence or absence of 20 mM MeAIB. Neurons were double-labeled for MAP2 and GABA, traced and analyzed using the Neurolucida software. Data are the mean Ϯ S.E. percentages of Ctrl values of 21-40 neurons from at least three independent experiments. *, p Ͻ 0.05, ***, p Ͻ 0.001 between Ctrl and BDNF-treated cultures, xx , p Ͻ 0.01 between BDNF-and BDNF ϩ MeAIB-treated cultures.  (Fig. 4, D and E). Immunocytochemical analysis of the cellular distribution of SNAT1 in cortical neurons revealed that, in agreement with previous findings (10,11), SNAT1 was expressed in both glutamatergic and GABAergic neurons, as 94.3 Ϯ 4.7% (n ϭ 1005 neurons from three independent experiments) of the cells were immunopositive for SNAT1.

% of Ctrl % of Ctrl
Role of SNAT1 in the Stimulation of Dendritic Development by BDNF-We next investigated the role of SNAT1 in the regulation of cortical dendritic development by BDNF. To address this issue, we used RNA interference to prevent the induction of SNAT1 expression by BDNF. A small interfering RNA targeted to mouse SNAT1 mRNA (SNAT1 siRNA) abolished SNAT1 protein expression in HEK293 cells transiently transfected with HA-SNAT1 plasmid (Fig. 5A). We next tested the efficacy of SNAT1 siRNA in cortical neurons. Western blot analysis showed that transfection of cortical neurons with SNAT1 siRNA markedly reduced SNAT1 protein levels in unstimulated (86 Ϯ 4.2% inhibition) and BDNF-stimulated cultures (73.7 Ϯ 5.8% inhibition), indicating that the selected SNAT1 siRNA was effective in decreasing endogenous and BDNF-induced SNAT1 expression in cortical neurons (Fig. 5, B and C). Biochemical studies revealed that application of SNAT1 siRNA to cortical neurons caused a significant reduction of BDNFinduced System A activity (49.2 Ϯ 8% inhibition, n ϭ 9) and protein synthesis (31.4 Ϯ 5.3% inhibition, n ϭ 17). The role of SNAT1 in the regulation of dendritic development by BDNF was tested by transfecting cortical neurons with SNAT1 siRNA together with pEGFP as a marker of transfected neurons. Although in cultures exposed to Ctrl siRNA, BDNF enhanced the dendritic length and branching of cortical neurons, treatment of cortical neurons with SNAT1 siRNA abolished BDNFinduced dendritic length and branching (Fig. 6). Suppression of BDNF-induced dendritic development by SNAT1 siRNA was not due to a reduction of neuronal viability, as measured by LDH release and MTT reduction (Fig. 3, C and D). Together, these results indicate that induction of SNAT1 expression is required for the stimulation of cortical dendritic development by BDNF.
Release of tPA Is Necessary for the Regulation of Dendritic Development by BDNF-Among proteins whose expression is induced by BDNF in developing cortical neurons, we have shown previously that BDNF increases the expression and secretion of tPA (12), a serine protease that regulates neurite outgrowth, neuronal migration, and synaptic plasticity (18 -20). Because there is evidence that tPA can regulate neurite outgrowth (18, 20), we examined the role of BDNF-induced tPA secretion in the regulation of dendritic morphology.
To address this issue, we tested the effect of 2,7-bis-(4-amidinobenzylidene)-cycloheptan-1-one (tPA stop), a bis-benzamidine derivative that  inhibits single-and two-chain forms of tPA (21), on the increased dendritic length and branching elicited by BDNF. Treatment of cortical neurons with tPA stop suppressed BDNF-induced dendritic length and branching (Fig. 7), indicating that secretion of tPA is required for the regulation of cortical dendritic development by BDNF. Role of System A and SNAT1 in the Up-regulation of tPA Levels by BDNF-On the basis of these observations, we determined whether stimulation of System A activity by BDNF was necessary for the elevation of tPA levels by BDNF. Zymographic analysis revealed that treatment of cortical neurons with MeAIB markedly reduced (65% inhibition) increases in tPA activity levels by BDNF (Fig. 8, A and B). These results indicate that up-regulation of System A amino acid transport plays a critical role in the enhancement of tPA levels by BDNF. We next examined whether induction of SNAT1 expression contributes to increases in tPA levels by BDNF. In agreement with the marked reduction of BDNF-induced tPA levels by MeAIB, treatment of cortical neurons with SNAT1 siRNA strongly reduced (69% inhibition) the increased tPA activity by BDNF (Fig. 8, C and D), providing evidence that induction of SNAT1 expression is essential for the enhancement of tPA levels by BDNF.

DISCUSSION
Dendritic development is guided by an intrinsic growth program that is capable of generating a basic dendritic arborization as well as by several extracellular factors such as neuronal activity, growth factors, and guidance molecules that are essential for sculpting dendrites to their final form (22)(23)(24). Of these extracellular factors, BDNF plays an important role in regulating the growth and branching of cortical dendrites (3,4). Regulation of dendritic growth by BDNF is thought to require the activation of a transcriptional program and new protein synthesis (22,24). Consistent with this hypothesis, BDNF increases both overall and local protein synthesis in cortical neurons (13,25,26). In this study, we examined the role of amino acid transport in the regulation of cortical dendritic development by BDNF. We show that BDNF increases amino acid transport mediated by System A (Fig. 1A), the major system responsible for the transport of short-chain neutral amino acids. Kinetic analysis of MeAIB uptake revealed that stimulation of System A   FEBRUARY 23, 2007 • VOLUME 282 • NUMBER 8 amino acid transport by BDNF is accompanied by an increase in the number of functional System A transporters in cortical neurons (Fig. 1, C and D). In agreement with this observation, BDNF enhances the expression of the System A transporter isoform SNAT1 (Fig. 4, B-E). Induction of SNAT1 expression by BDNF together with the inhibition of BDNF-induced MeAIB uptake by SNAT1 siRNA suggest that BDNF increases System A activity by up-regulating SNAT1 expression. It is noteworthy that a statistically significant effect of BDNF on System A activity is observed after 20 min of stimulation by BDNF (Fig. 1A), whereas SNAT1 protein expression is not increased (not shown). This rapid elevation of System A activity by BDNF may result from the translocation of SNAT1 transporters from an intracellular compartment to the plasma membrane of cortical neurons. In support of this hypothesis is the recent observation that the cellular distribution of SNAT1 immunoreactivity in cortical neurons appears punctate, suggesting the presence of large intracellular pools of SNAT1 transporters (9,10).

BDNF regulates System A Amino Acid Transport
Our results show that stimulation of SNAT1 expression and System A amino acid transport by BDNF plays a major role in the regulation of cortical dendritic development by BDNF. Thus, inhibition of BDNF-induced SNAT1 expression and System A activity by siRNA and MeAIB, respectively, abolishes the increased dendritic growth and branching of cortical neurons by BDNF (Figs. 2 and 6). Together, these data reveal a novel mechanism by which BDNF stimulates amino acid transport mediated by System A to meet the increased metabolic demand associated with the morphological differentiation of cortical neurons by BDNF.
Among proteins underlying dendritic development, there is evidence that the serine protease tPA is secreted from growing neurites and promotes neurite outgrowth (18,20). With regard to the regulation of tPA, we have shown previously that BDNF enhances the expression and release of tPA in developing cortical neurons (12). Here, we provide strong evidence that increases in tPA levels by BDNF require up-regulation of SNAT1 expression and System A amino acid transport. In support of this observation, the enhancement of tPA levels by BDNF is strongly reduced by suppressing BDNF-induced SNAT1 expression or System A activity (Fig. 8). Our results also demonstrate that tPA stop blocks the increased dendritic length and branching of cortical neurons by BDNF (Fig. 7), indicating that secretion of tPA by BDNF is essential for the effects of BDNF on dendritic development. Together, these data provide an example of a protein whose induction by BDNF requires the up-regulation of SNAT1 expression and System A activity and whose secretion plays a critical role in the regulation of cortical dendritic development by BDNF.
In addition to the regulation of cortical dendritic morphology during development, data presented in this study could be relevant to cortical plasticity, in particular to ocular dominance plasticity. In support of this hypothesis, BDNF and tPA have been identified as key regulators of ocular dominance plasticity (27)(28)(29)(30)(31). For instance, BDNF regulates the critical period for ocular dominance plasticity (27), and tPA plays a permissive role for visual cortical plasticity during the developmental critical period (29 -31). Moreover, ocular dominance plasticity requires cortical protein synthesis as the protein synthesis inhibitor cycloheximide blocks the effects of monocular deprivation (32). Our data suggest that up-regulation of SNAT1 expression and System A amino acid transport by BDNF may represent an important mechanism underlying structural remodeling during the critical period for ocular dominance plasticity. Besides visual cortical plasticity, BDNF is involved in hippocampal long-term potentiation (LTP). Thus, LTP is impaired markedly in BDNF knock-out mice (33,34), whereas reexpression of BDNF in the CA1 region (33) or treatment with recombinant BDNF restores LTP (34). There is also compelling evidence that the late phase LTP that persists for at least a day requires new protein synthesis (35). In this regard, it is of particular interest that tPA is induced by LTP (36) and contributes to the late phase LTP (37,38). Together, these observations suggest that new protein synthesis is necessary for the effects of BDNF during the late phase LTP. In line with this hypothesis, our findings imply that stimulation of SNAT1 expression and System A activity by BDNF may play an important role in the synthesis of new proteins associated with the late phase LTP by supplying amino acids for the synthesis of effectors such as tPA.
In conclusion, our data indicate that promotion of neuronal differentiation by BDNF requires the up-regulation of SNAT1 expression and System A amino acid transport to ensure an adequate substrate supply for the increased synthesis of proteins necessary for the enhancement of dendritic growth and branching. These results underscore the importance of the regulation of amino acid transport in the control of dendritic development.