Skip to main content
NCBI home page
As a library, NLM provides access to scientific literature. Inclusion in an NLM database does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health.
Learn more: PMC Disclaimer | PMC Copyright Notice
[画像:Proceedings of the National Academy of Sciences of the United States of America logo]
. 2009 Jul 10;106(30):12524–12529. doi: 10.1073/pnas.0903328106

Glial ephrin-A3 regulates hippocampal dendritic spine morphology and glutamate transport

Maria A Carmona a, Keith K Murai b, Lei Wang a, Amanda J Roberts c, Elena B Pasquale a,d,1
aBurnham Institute for Medical Research, 10901 North Torrey Pines Road, La Jolla, CA 92037;
bDepartment of Neurology and Neurosurgery, Centre for Research in Neuroscience, McGill University Health Centre, Montreal General Hospital, 1650 Cedar Avenue, Montreal, QC, Canada H3G 1A4;
cMolecular and Integrative Neurosciences Department, The Scripps Research Institute, La Jolla, CA 92037; and
dPathology Department, University of California San Diego, La Jolla, CA 92093
1

To whom correspondence should be addressed. E-mail: elenap@burnham.org

Edited by Anthony J. Pawson, University of Toronto, Toronto, Canada, and approved May 28, 2009

Author contributions: M.A.C., K.K.M., and E.B.P. designed research; M.A.C., L.W., and A.J.R. performed research; K.K.M. contributed new reagents/analytic tools; M.A.C. and A.J.R. analyzed data; and M.A.C., K.K.M., and E.B.P. wrote the paper.

Received 2009 Mar 26; Issue date 2009 Jul 28.

PMCID: PMC2718351 PMID: 19592509

Abstract

Increasing evidence indicates the importance of neuron-glia communication for synaptic function, but the mechanisms involved are not fully understood. We reported that the EphA4 receptor tyrosine kinase is in dendritic spines of pyramidal neurons of the adult hippocampus and regulates spine morphology. We now show that the ephrin-A3 ligand, which is located in the perisynaptic processes of astrocytes, is essential for maintaining EphA4 activation and normal spine morphology in vivo. Ephrin-A3-knockout mice have spine irregularities similar to those observed in EphA4-knockout mice. Remarkably, loss of ephrin-A3 or EphA4 increases the expression of glial glutamate transporters. Consistent with this, glutamate transport is elevated in ephrin-A3-null hippocampal slices whereas Eph-dependent stimulation of ephrin-A3 signaling inhibits glutamate transport. Furthermore, some forms of hippocampus-dependent learning are impaired in the ephrin-A3-knockout mice. Our results suggest that the interaction between neuronal EphA4 and glial ephrin-A3 bidirectionally controls synapse morphology and glial glutamate transport, ultimately regulating hippocampal function.

Keywords: glial glutamate transporters, hippocampus-dependent learning, neuron-glia communication, perisynaptic glial processes

Communication between neurons and astrocytes plays an important role in synapse development and physiology. During development, astrocytes promote synapse formation and maturation (1, 2). Later on, astrocytes are actively involved in synaptic transmission by responding to synaptic activation and releasing gliotransmitters, which in turn modulate neuronal excitability and neurotransmission (1, 3). At excitatory synapses, astrocytes clear the majority of glutamate released into the synaptic cleft through their high-affinity transporters, GLAST and GLT-1 (4). This function is crucial for fine-tuning glutamatergic transmission and prevents accumulation of toxic levels of extracellular glutamate (5–7). Despite their importance in glutamate homeostasis, very little is known about the molecular mechanisms that regulate the expression and cell surface localization of glial glutamate transporters.

Several members of the Eph family of receptor tyrosine kinases (comprising EphA and EphB receptors) and their cell surface-associated ephrin ligands are expressed in neurons of the hippocampus and cerebral cortex, where they have been implicated in synapse formation and the regulation of synaptic function and plasticity (8, 9). Signaling by Eph receptors is also involved in the development of dendritic spines, the specialized protrusions on dendrites that receive excitatory innervation. Typical spines have an enlarged head connected to the dendritic shaft by a constricted neck. Activation of dendritic EphB receptors by ephrin-B ligands has been shown to induce spine morphogenesis and maturation (8, 9). EphA receptors are important regulators of spine morphology in the adult. We have reported that EphA4 is localized on dendritic spines of pyramidal neurons in the mature hippocampus (10). Activation of EphA4 decreases spine length and density whereas loss of EphA4 activity results in spine elongation and disorganization. The major ephrin-A ligand for EphA4 present in the adult hippocampus is ephrin-A3, which is expressed in astrocytes (10, 11). However, whether glial ephrin-A3 affects spine morphology remains to be clarified.

A notable feature of Eph receptors and ephrins is that their interaction triggers bidirectional signals between the engaged cells (8). Forward signals are mediated by the Eph receptors and reverse signals by the ephrins. Signaling pathways downstream of EphA4 that induce changes in spine shape and density have been recently elucidated (8, 9). Reverse signals downstream of A-type ephrins, which are GPI-linked proteins, have also been reported in some instances (8, 9). However, the importance of ephrin-A reverse signaling in vivo has not been well characterized.

By analyzing ephrin-A3-knockout mice, we now show that this ephrin is a critical partner for EphA4 in the regulation of hippocampal dendritic spine morphology. Importantly, ephrin-A3 and EphA4 also modulate the expression of glial glutamate transporters and glutamate transport. Taken together, our results suggest a mechanism regulating hippocampal function through neuron-glia interaction via EphA4 and ephrin-A3.

Results

Ephrin-A3-Knockout Mice Have Abnormalities in Hippocampal Dendritic Spines.

EphA4 receptor signaling regulates dendritic spine morphology in hippocampal pyramidal neurons (8–10). To examine the importance of ephrin-A3 for this regulation, we inactivated the ephrin-A3 gene in the mouse (Fig. S1 a–e). Overall hippocampal cytoarchitecture and the dendritic arborization of diolistically labeled CA1 pyramidal neurons appeared normal in the knockout mice (Fig. S1 f and g). The spines, however, had a disorganized appearance and showed significant elongation in 7-week- and 8-month-old ephrin-A3-knockout mice, with significantly longer spine necks (Fig. 1 A–E and Fig. S2 a–e). This phenotype is reminiscent of that observed in EphA4-knockout mice (10). Furthermore, classification of spine morphologies revealed a gain in mushroom-shaped spines and a loss of stubby spines in ephrin-A3-null CA1 pyramidal neurons (Fig. 1 F and Fig. S2f). Spine density and width were not significantly affected by the lack of ephrin-A3 (Fig. 1 A–C and Fig. S2 a–c). These abnormalities indicate that ephrin-A3 regulates spine structure in vivo, and imply that its effects are not compensated during adulthood or by other EphA4 ligands.

Fig. 1.

Fig. 1.

Abnormal in vivo dendritic spine morphology in ephrin-A3-knockout mice. (A) Three-dimensional reconstruction of diolistically labeled dendritic spines from CA1 pyramidal neurons of 7-week-old wild-type (+/+) and ephrin-A3-knockout (−/−) mice. Arrows point to examples of mushroom (m), stubby (s), or thin (t) spines. (B–E) Quantification of dendritic spine parameters. The ephrin-A3 knockout mice have similar spine density and width compared with control (+/+ and +/−) mice, but their spines are significantly longer and have longer spine necks (***, P < 0.001, KS and Student's t test). (F) Ephrin-A3-knockout mice have more mushroom-shaped spines and fewer stubby spines than control mice (***, P < 0.001, one-way ANOVA with Bonferroni's post hoc comparisons). Error bars, SEM. (Scale bar, 1 μm.)

EphA4 Activation Is Reduced in the Ephrin-A3-Null Hippocampus.

Using an antibody that specifically detects EphA4, and does not label the brain of EphA4-knockout mice, we did not observe changes in the expression and distribution of EphA4 in the ephrin-A3-null hippocampus (Fig. S3 a–d). Double labeling showed an extensive colocalization of EphA4 and PSD-95 in the hippocampus of wild-type and ephrin-A3-knockout mice (Fig. S3 e–j), in agreement with the postsynaptic localization of EphA4 (10). Remarkably, we detected a 50% reduction in EphA4 tyrosine phosphorylation, indicating a strong decrease in receptor activation (Fig. 2 A and B). The remaining EphA4 receptor phosphorylation observed in the absence of ephrin-A3 may be due to activation by other ephrins, such as B-ephrins (9), which are expressed at normal levels in the knockout mice (Fig. 2 C). Hence, endogenous ephrin-A3 is a major ligand for EphA4 in the adult hippocampus and is needed for proper spine morphology in vivo.

Fig. 2.

Fig. 2.

Ephrin-A3 promotes EphA4 phosphorylation in the adult hippocampus. (A) Immunoprecipitation and immunoblot analysis indicates lower EphA4 phosphorylation in the 6- to 8-week-old ephrin-A3-null hippocampus (−/−) compared with wild type (+/+). (B) Quantification of 3 experiments shows that EphA4 phosphorylation in ephrin-A3 knockout mice is ≈50% that in wild-type mice (*, P < 0.05, Student's t test). Error bar, SEM. (C) The expression levels of B-type ephrins, analyzed by immunoblotting after immunoprecipitation with a pan-ephrin-B antibody, are similar in wild-type and ephrin-A3-null hippocampus.

To determine whether EphA4 signaling can take place in the ephrin-A3-knockout mice, we activated EphA4 in hippocampal slices. Stimulation with a soluble dimeric form of ephrin-A3 (ephrin-A3 Fc) induced similar EphA4 activation and spine retraction in acute hippocampal slices from ephrin-A3 knockout and wild-type mice, demonstrating that EphA4 signaling ability is normal in pyramidal neurons of the ephrin-A3-null mice (Fig. S4).

Ephrin-A3 Colocalizes with GLAST in Astrocytic Processes.

Previously it was shown that ephrin-A3 mRNA is expressed in astrocytes of the adult mouse hippocampus (10, 11). Consistent with this, double labeling with an antibody specific for ephrin-A3 (Fig. S1d) showed that ephrin-A3 extensively colocalizes with puncta positive for the glial-specific glutamate/aspartate transporter (GLAST; Fig. 3 A), which is concentrated in the perisynaptic processes of astrocytes (12). Furthermore, ephrin-A3 did not show overlapping expression with excitatory presynaptic terminals labeled for the vesicular glutamate transporter 1 (VGLUT-1) (Fig. 3 B) or with postsynaptic structures labeled for the postsynaptic density protein, PSD-95 (Fig. 3 C). Instead, ephrin-A3 immunostaining surrounded VGLUT-1- and PSD-95-positive puncta, in agreement with its previously reported localization juxtaposed to synapses (10). The pattern of ephrin-A3 labeling appeared to be developmentally regulated, because ephrin-A3 was found in scattered glial fibrillary acidic protein (GFAP)-positive somata through the hippocampal neuropil during early postnatal development (Fig. 3 D and F–H) before establishing its mature pattern of expression in astrocytic processes that flank synapses (Fig. 3 B, C, and E).

Fig. 3.

Fig. 3.

Ephrin-A3 colocalizes with GLAST on astrocytic processes in the adult hippocampus. (A–C) Immunofluorescence confocal images of hippocampal sections from adult mice double-labeled for ephrin-A3 (green) and GLAST (red, A), VGLUT-1 (red, B), and PSD-95 (red, C). Arrows in B and C show examples where ephrin-A3 labeling surrounds presynaptic (B) and postsynaptic (C) glutamatergic terminals. Extensive colocalization with GLAST in A indicates that ephrin-A3 is concentrated in perisynaptic astrocytic processes. (D and E) Overview of ephrin-A3 expression in the hippocampus of P10 and P21 mice. (F–H) Double labeling for ephrin-A3 (F) and GFAP (G) in P10 hippocampus shows extensive colocalization in the merged image (H) (arrows). DG, dentate gyrus; slm, stratum lacunosum moleculare; sr, stratum radiatum. [Scale bars, 3 μm (C); 50 μm (D and E); and 20 μm (H).]

Glial Glutamate Transporters Are Up-Regulated in the Ephrin-A3 and the EphA4-Null Hippocampus.

Because Eph/ephrin complexes can transmit bidirectional signals, and ephrin-A3 colocalizes with GLAST, we investigated whether loss of ephrin-A3 affects expression of GLAST/EAAT1. We also examined glutamate transporter-1 (GLT-1/EEAT2), which is the most abundant glutamate transporter present in perisynaptic processes of hippocampal astrocytes (4, 6, 12). Strikingly, quantitative analysis of confocal images showed marked up-regulation of GLT-1 and GLAST in the ephrin-A3-null hippocampus (Fig. 4 A–C). Interestingly, an elevated percentage of bright pixels (Fig. 4 C) (see Materials and Methods) suggests an increase of GLT-1 and GLAST expression in individual astrocytic processes. The levels of GFAP as well as the density and distribution of astrocytes were not altered in the ephrin-A3 null hippocampus (Fig. S5). Expression of the neuronal glutamate transporter excitatory amino acid carrier 1 (EAAC1/EAAT3) was also not altered (Fig. 4 A–C), indicating selective changes in the glial glutamate transporters. Immunoblot analysis of hippocampal lysates confirmed elevated overall expression of GLT-1 and GLAST, but not EAAC1 (Fig. 4 F and G) or other synaptic proteins (Fig. S6).

Fig. 4.

Fig. 4.

Enhanced levels of glial glutamate transporters in ephrin-A3 and EphA4-knockout mice. (A) Single-plane confocal images showing immunofluorescence labeling for GLT-1, GLAST, and EAAC1 in the CA1 stratum radiatum of the wild-type and ephrin-A3-null hippocampus. (Scale bar, 8 μm.) (B) Quantification of average pixel intensities reveal higher immunoreactivities for GLT-1 and GLAST but not EEAC1 in the ephrin-A3-null hippocampus. (C) The percentage of bright pixels (corresponding to the 5% brightest pixels in the wild-type sections and the pixels as bright or brighter in the knockout sections) indicates higher labeling intensity for GLT-1- and GLAST-positive punctae in the ephrin-A3-knockout mice. (D) Quantification of average pixel intensities in the EphA4-null hippocampus relative to wild type. Average pixel intensities are significantly higher only for GLT-1 in the EphA4-knockout mice compared with wild type. (E) The percentage of bright pixels in the EphA4-knockout hippocampus is higher than wild type for GLT-1, GLAST, and EAAC1, but not ephrin-A3. (F) Immunoblots comparing expression of the indicated proteins in wild-type and ephrin-A3-null hippocampus. (G) Quantification of immunoblots from 4 pairs of wild-type and ephrin-A3-knockout mice by densitometry indicates higher expression of glial transporters in the knockout hippocampus whereas expression of the neuronal glutamate transporter EAAC1 is similar. Values were normalized to GAPDH levels and expressed as a percentage of the levels in the wild-type hippocampus. (H) mRNA levels are similar in the wild-type and ephrin-A3-null hippocampus, as determined by real-time RT-PCR analysis. Values were normalized to GAPDH transcript levels and expressed as a percentage of the levels in wild-type hippocampus. Error bars, SEM; *, P < 0.05, **, P < 0.01, ***, P < 0.001, Student's t test.

Up-regulation of GLT-1 and GLAST was also observed in the hippocampus of EphA4 knockout mice (Fig. 4 D and E). However, in addition, expression of EAAC1 was affected by the loss of EphA4 (Fig. 4 E). Expression of ephrin-A3 was not altered in the EphA4-knockout mice. Thus, the interaction between ephrin-A3 and EphA4 maintains normal GLT-1 and GLAST expression in hippocampal astrocytes. Only modest or nonsignificant changes in GLT-1 and GLAST expression were observed in several other brain regions, where endogenous ephrin-A3 and/or EphA4 are present at significantly lower levels than in the hippocampus (Fig. S7). These data support a role for ephrin-A3 and EphA4 in limiting glial glutamate transporter levels.

Double-labeling experiments revealed a complementary expression of EphA4 and GLAST (Fig. S8), in agreement with previous reports showing minor, if any, EphA4 expression in astrocytes of the normal adult brain (10, 13). Therefore, the interaction of neuronal EphA4 with astrocytic ephrin-A3 down-regulates glial glutamate transporters. Real-time qPCR showed similar GLAST and GLT-1 mRNA levels in the ephrin-A3-null and wild-type hippocampus, indicating that the down-regulation likely occurs at the posttranscriptional level (Fig. 4 H).

Ephrin-A3 Reverse Signaling Inhibits Glutamate Transport.

We next investigated whether the elevated levels of GLT-1 and GLAST in the ephrin-A3-knockout hippocampus correlate with more efficient glutamate uptake. By using L-[3H]-glutamate as a radioactive tracer, we measured 50% higher uptake in ephrin-A3-null acute hippocampal slices compared with wild type (Fig. 5 A). Glutamate uptake was reduced to similar low levels in wild-type and ephrin-A3-null slices treated with threo-β-benzyloxyaspartate (TBOA) or in the absence of extracellular Na (Fig. 5 A), consistent with the fact that high-affinity glutamate transporters on both neurons and glia are inhibited by TBOA and require extracellular Na for activity (4). Dihydrokainate (DHK), a selective GLT-1 inhibitor (4), decreased glutamate uptake by 35% in wild-type slices (mean ± SEM: basal, 1275 ± 75 fmol/μg/min; DHK, 830 ± 93 fmol/μg/min), in agreement with previous studies (14), and by 42% in ephrin-A3-null slices (mean ± SEM: basal, 1909 ± 35 fmol/μg/min; DHK, 1104 ± 121 fmol/μg/min) (Fig. 5 A). This corresponds to a decrease of 445 and 805 fmol/μg/min in wild-type and ephrin-A3-null slices, respectively, which represents 80% more DHK-sensitive uptake in the knockout slices. Thus, the elevated uptake in ephrin-A3-null slices is mostly due to increased function of GLT-1, which is the major glutamate transporter in the adult brain (5).

Fig. 5.

Fig. 5.

Ephrin-A3 regulates glutamate uptake. (A) Na+-dependent uptake is 50% higher in ephrin-A3-null acute hippocampal slices (***, P < 0.001, one-way ANOVA with Bonferroni post hoc comparisons). The general glutamate transporter inhibitor TBOA and absence of extracellular Na+ reduce uptake to similar basal levels in ephrin-A3-null and wild-type slices; the GLT-1 inhibitor DHK also reduces uptake and eliminates the difference in uptake between ephrin-A3-null and wild-type slices (basal versus DHK: P < 0.05 for wild-type and P < 0.001 for knockout; basal versus TBOA, P < 0.001 for wild-type and P < 0.001 for knockout; one-way ANOVA with Bonferroni post hoc comparisons). (B) Stimulation of hippocampal slices with EphA2 Fc at room temperature for 10 h (RT) or at 32 °C for 2 h (32 °C) decreases glutamate uptake in wild-type but not ephrin-A3-null slices. Heat inactivated EphA2 Fc is ineffective. Percentage changes in glutamate uptake are determined relative to slices treated with Fc from the same experiment. Error bars, SEM. **, P < 0.01 by Student's t test for the comparison with Fc control. (C and D) Stimulation of wild-type hippocampal slices at room temperature for 10 h down-regulates GLT-1 and GLAST levels. (C) Representative examples of immunoblots of hippocampal slices stimulated with Fc or EphA2 Fc. (D) Densitometric quantification of 3 independent experiments shows lower expression of GLT-1 and GLAST in EphA2 Fc-stimulated slices (**, P < 0.01, Student's t test), compared with Fc whereas expression of GFAP is similar. Values were normalized to GAPDH levels and expressed as a percentage of the levels in Fc-treated slices. Error bars, SEM.

We also investigated the involvement of ephrin-A3 reverse signaling in glutamate transport by stimulating hippocampal slices with EphA2 Fc, which selectively binds to A-type ephrins (whereas EphA4 Fc can also bind to B-type ephrins). EphA2 Fc caused a marked decrease in glutamate uptake in wild-type but not ephrin-A3-null slices (Fig. 5 B). Consistent with this, EphA2 Fc also reduced GLT-1 and GLAST levels in wild-type slices (Fig. 5 C and D). Thus, EphA binding to ephrin-A3 down-regulates glutamate transporters and glutamate uptake.

The Ephrin-A3-Knockout Mice Have Defects in Hippocampus-Dependent Learning.

General tests for behavior indicated that locomotor activity, sensorimotor responses, motor coordination, anxiety, and depression are similar in ephrin-A3-knockout and wild-type mice (Fig. S9). To more selectively probe for deficits in hippocampal-dependent learning, we used a variety of behavioral tasks. In the fear conditioning test, acquisition of cue-associated fear memory (association of a stimulus, such an acoustic tone, with a mild footshock) depends on the amygdala whereas acquisition of contextual fear memory (association of contextual cues with the footshock) also depends on the hippocampus (15). The time spent "freezing" is measured as an indication of fear. The freezing responses to the acoustic tone were similar between wild-type and ephrin-A3-knockout mice, indicating normal cue-associated memory (Fig. 6 A). However, the freezing responses to the contextual cues were significantly reduced in the ephrin-A3-knockout mice (Fig. 6 A), indicating impaired acquisition of hippocampus-dependent contextual fear memory.

Fig. 6.

Fig. 6.

Learning and memory performance in ephrin-A3-knockout mice. (A) Fear conditioning test. Based on freezing responses, the ephrin-A3-knockout mice showed a significantly impaired context-associated fear memory compared with wild-type mice (***, P < 0.001, one-way ANOVA). Freezing responses to the acoustic conditioned stimuli (CS) were similar in wild-type and ephrin-A3-knockout mice, indicating normal cue-associated fear memory. (B and C) Object placement tests. (B) Wild-type mice, but not ephrin-A3-knockout mice, spent significantly more time near the object placed in a new location (*, P < 0.01, ANOVA with paired t test), indicating that only wild-type mice remembered the previous location of the object. (C) This difference was abolished by habituating the mice to the context for 30 min before introducing the object. (D) Barnes maze test. The ephrin-A3-knockout mice made a similar number of errors as the wild-type mice, indicating normal acquisition of spatial memory in this test (P = 0.5, one-way ANOVA). Error bars, SEM.

In the object placement test, which is also known to depend on the hippocampus (16), the mice become familiar with the location of an object in a test chamber and then show renewed interest when the object is moved. During the training periods, the ephrin-A3-knockout mice spent more time exploring the object than wild-type mice. However, the wild-type mice but not ephrin-A3-knockout mice showed renewed interest in the object when the object was moved (Fig. 6 B), indicating that loss ephrin-A3 impairs memory of the previous location. Intriguingly, in a variation of the test in which the mice were allowed to explore the test chamber for 30 min before object introduction, the ephrin-A3-knockout mice did not show significant deficiencies (Fig. 6 C). Therefore, the knockout mice exhibited deficiencies only in the version of the object placement test in which exposure to the context was brief, and the context and object had to be assimilated at the same time. This suggests deficiencies in memorizing contextual information, which is consistent with the fear conditioning results.

In the Barnes maze, which assesses hippocampus-dependent spatial memory, the mice learn to find a hidden exit situated in a constant location by using visual cues and are repeatedly tested over a period of many days. The number of errors in finding the hidden exit was comparable in wild-type and ephrin-A3-knockout mice (Fig. 6 D), indicating similar spatial memory performance.

Discussion

Our findings show that the interaction between EphA4 and ephrin-A3 at hippocampal excitatory synapses leads to bidirectional effects in neurons and glial cells (Fig. 7). The EphA4 receptor located on dendritic spines modulates spine morphology upon binding to ephrin-A3 whereas activation of astrocytic ephrin-A3 by neuronal EphA4 regulates glial glutamate transporter expression and glutamate transport. Our results reveal a mechanism by which glial glutamate transport is controlled near synapses via EphA4 and ephrin-A3. Although EphA4 is expressed both presynaptically and postsynaptically (13), the effects of ephrin-A3 on spine morphology reported here and the fact that glial coverage in hippocampal synapses is asymmetric and preferentially involves dendritic spines (17), suggest a preferential interaction of ephrin-A3 with postsynaptic EphA4. Moreover, the segregation of EphA4 and GLAST immunoreactivities reported here agrees with the separate localizations of EphA4 and ephrin-A3 (10) and with the notion that EphA4 is not widely expressed in adult astrocytes in vivo (13).

Fig. 7.

Fig. 7.

EphA4/ephrin-A3 bidirectional effects at hippocampal synapses. Binding of glial ephrin-A3 to postsynaptic EphA4 activates EphA4 forward signaling in neurons, which induces dendritic spine retraction, and ephrin-A3 reverse signaling in astrocytes, which down-regulates glutamate transport. The presynaptic terminal is depicted in gray, the postsynaptic dendritic spine is in red, and a glial cell process near the synapse is in blue.

We reported that activation of EphA4 induces dendritic spine shortening and retraction, and that spines are longer than normal in the hippocampus of EphA4-knockout mice (10). We now show that EphA4 phosphorylation is reduced in the ephrin-A3-null hippocampus, concomitant with an abnormal spine elongation. Thus, absence of either EphA4 or ephrin-A3 causes equivalent spine phenotypes, consistent with a role for the EphA4/ephrin-A3 system in maintaining postsynaptic morphology. Studies have shown an association between spine structure and synapse physiology (18–20). In particular, neck length is one of the factors that control the rate of decay of calcium signals, because they diffuse from the synapse to the dendrite. Therefore, the increased length of the spine necks observed in the ephrin-A3-knockout mice might alter calcium dynamics in spines, which in turn has been related to learning. Consistent with this, we show that loss of ephrin-A3 causes impairments in contextual fear memory and in the object placement test, 2 paradigms of learning and memory that are dependent on the hippocampus, supporting a role for glial ephrin-A3 in cognitive processes.

We describe a role for ephrin-A3 in maintaining physiological levels of glial glutamate transporters through contact with EphA4. The loss of ephrin-A3 or EphA4 increases GLT-1 and GLAST protein expression in the hippocampus, but not mRNA expression, indicating a posttranscriptional form of regulation. In line with this, we found that stimulation of acute hippocampal slices with EphA2 Fc reduces glial glutamate transporter levels and glutamate uptake, arguing that acute signaling by ephrin-A3 may lead to the degradation of glial glutamate transporters and a consequent decrease of glutamate uptake. This is consistent with the higher levels of glutamate transporters and glutamate uptake observed in the hippocampus of the ephrin-A3-knockout mice.

The signals downstream of ephrin-A3 that reduce glial glutamate transporter expression are currently unknown. Both glial glutamate transporters and ephrin-A ligands are associated with cholesterol-rich lipid rafts microdomains of the plasma membrane (20, 21). Contact with EphA4 may drive endocytosis of ephrin-A3-containing plasma membrane vesicles, as reported for other ephrins (22), resulting in the concomitant internalization of the colocalized transporters and their subsequent degradation. Internalization and degradation of the transporters may also be caused by intracellular signaling pathways activated by ephrin-A3. This may occur through assembly of signaling complexes triggered by clustering of the GPI-linked ephrin-A3 in lipid rafts and/or by association of ephrin-A3 with a transmembrane protein (8, 9). Interestingly, protein kinase C activation has been shown to down-regulate surface expression and activity of GLT-1 and to induce its ubiquitination and degradation (23).

Cross-talk between astrocytes and neurons mediated by diffusible factors has been shown to up-regulate GLT-1 and GLAST mRNA and protein expression in cultured astrocytes (4). We now show that contact-dependent neuron-astrocyte interaction down-regulates GLT-1 and GLAST at the posttranscriptional level. The combination of these positive and negative signals may be responsible for the ability of glial glutamate transporters to dynamically adjust to synaptic activation and sensory stimulation (7).

The increased levels and function of glial glutamate transporters that occur in the ephrin-A3-knockout mice suggest decreased glutamate concentrations near synapses. This may affect the extent of postsynaptic activation and reduce spillover of glutamate, thus altering neuronal transmission and plasticity (7). Interestingly, the EphA4 knockout mice show deficits in LTP that are independent of forward signaling (9). These defects might be due, at least in part, to deficient ephrin-A3 reverse signaling.

The ephrin-A3-knockout mice show impairments in certain learning and memory tasks that require the hippocampus. Contextual fear memory is impaired whereas spatial memory (assessed using the Barnes maze) is normal. Even though both require the hippocampus (15, 24), contextual and spatial memories might be encoded by different mechanisms and circuits, and have been reported to be independently affected in other mouse models (24–26). The fact that the ephrin-A3-knockout mice perform normally in the prehabituation version of the object placement test, but fail in a more demanding version of the test in which both contextual (test chamber) and spatial (object location) information have to be assimilated simultaneously, might indicate that ephrin-A3-knockout mice have moderate hippocampal deficiencies, which are revealed only when the mice have to perform more cognitively demanding tasks.

Our findings have pathological implications. Dendritic spine abnormalities accompany neurological disorders such as mental retardation, autism and schizophrenia (27). Increased glial glutamate transporter levels and function are also associated with schizophrenia (4, 28). However, dysfunction of glial glutamate transporters has been implicated in epilepsy and ALS (4, 6, 28). Increased EphA4-ephrin interaction may therefore underlie some forms of ALS and the posttranscriptional down-regulation of GLT-1 observed in epileptic foci of the human brain (29–31). Our observations suggest that treatments to inhibit ephrin-A3 reverse signaling may increase glutamate transporter expression and have protective effects against epileptic seizures and glutamate excitotoxicity.

Materials and Methods

Densitometric Analysis and Pixel Counts.

Immunoblotting experiments were quantified using National Institutes of Health ImageJ software. Optical density values were normalized to either GAPDH or actin signals. In the immunofluorescence experiments, average pixel intensities were determined using the "histogram" tool in Adobe Photoshop. For analysis of the percentage of bright pixels, the intensities of the 5% brightest pixels (in a scale from 0 black to 255 white) was determined for images of wild-type sections. The percentage pixels with the same or higher intensity values was calculated for images of ephrin-A3-knockout sections (n = 7–13 fluorescent images per experiment in which 3 wild-type and 3 knockout mice were analyzed).

Glutamate Uptake.

To measure glutamate uptake, transverse acute hippocampal slices (300-μm thickness) were equilibrated for 1 h at room temperature in oxygenated artificial cerebrospinal fluid (ACSF, 126 mM NaCl, 2.5 mM KCl, 1.2 mM NaH2PO4, 1.3 mM MgCl2, 2.4 mM CaCl2, 26 mM NaHCO3, and 10 mM glucose), incubated for 7 min in ACSF containing 9.5 nM L-[3H]-glutamate (0.5 μCi) and 100 μM unlabeled glutamate, and then rapidly chilled on ice (14). Slices were then rinsed with cold ACSF and solubilized in 1% Triton X-100. The radioactivity in 25 μL lysates was measured in a scintillation counter (Beckman LS 6000SC). Na+-independent L-[3H]-glutamate uptake was measured on ice using ACSF containing choline chloride instead of NaCl. The protein concentration of the samples was determined with the Bradford assay (Bio-Rad) and L-[3H]-glutamate uptake was normalized to protein content. In Fig. 5 A, 7 wild-type and 7 ephrin-A3 knockout mice were used in 7 independent experiments, each including 2 to 4 hippocampal sections per condition. To block DHK- and TBOA-sensitive L-[3H]-glutamate uptake, hippocampal slices were incubated with 500 μM DHK or 300 μM TBOA (Sigma–Aldrich) for 30 min before measuring uptake. The effects of DHK and TBOA were tested in 3 wild-type and 3 ephrin-A3-knockout mice in 3 independent experiments, each including 4 hippocampal slices per condition. In Fig. 5 B, wild-type and ephrin-A3-knockout hippocampal slices were stimulated with 10 μg/mL EphA2 Fc, or Fc as a control, in ACSF at room temperature (RT) for 10 h, or at 32 °C for 2 h, before measuring glutamate uptake. When heat-inactivated EphA2 Fc was used, EphA2 Fc was heated for 30 min at 95 °C, and was allowed to cool down before being used for stimulation. Five wild-type and 2 knockout mice were used for EphA2 Fc stimulation at RT, 2 wild-type mice were used for stimulation with heat-inactivated EphA2 Fc at RT, and 3 wild-type mice for stimulation with EphA2 Fc at 32 °C. Each experiment included 4 hippocampal sections per condition.

For detailed description of experimental procedures, see SI Text.

Supplementary Material

Supporting Information

Acknowledgments.

We thank F. Beltran and J. Chang for technical assistance, V. Sharma and H. Freeze for help with the glutamate uptake assays, and D. Feldheim (University of California, Santa Cruz, CA) for providing ephrin-A3 knockout brains from a different line for comparative immunohistochemistry experiments. This work was supported by National Institutes of Health Grant HD025938, Sanford Children's Health Research Center grant (to E.B.P.), and a postdoctoral fellowship from the Fundación Española para la Ciencia y la Tecnología (to M.A.C.).

Footnotes

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

This article contains supporting information online at www.pnas.org/cgi/content/full/0903328106/DCSupplemental.

References

  • 1.Allen NJ, Barres BA. Signaling between glia and neurons: Focus on synaptic plasticity. Curr Opin Neurobiol. 2005;15:542–548. doi: 10.1016/j.conb.200508006. [DOI] [PubMed] [Google Scholar]
  • 2.Nishida H, Okabe S. Direct astrocytic contacts regulate local maturation of dendritic spines. J Neurosci. 2007;27:331–340. doi: 10.1523/JNEUROSCI.4466-06.2007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Perea G, Araque A. Astrocytes potentiate transmitter release at single hippocampal synapses. Science. 2007;317:1083–1086. doi: 10.1126/science.1144640. [DOI] [PubMed] [Google Scholar]
  • 4.Danbolt NC. Glutamate uptake. Prog Neurobiol. 2001;65:1–105. doi: 10.1016/s0301-0082(00)00067-8. [DOI] [PubMed] [Google Scholar]
  • 5.Tanaka K, et al. Epilepsy and exacerbation of brain injury in mice lacking the glutamate transporter GLT-1. Science. 1997;276:1699–1702. doi: 10.1126/science.276.5319.1699. [DOI] [PubMed] [Google Scholar]
  • 6.Maragakis NJ, Rothstein JD. Glutamate transporters: Animal models to neurologic disease. Neurobiol Dis. 2004;15:461–473. doi: 10.1016/j.nbd.200312007. [DOI] [PubMed] [Google Scholar]
  • 7.Tzingounis AV, Wadiche JI. Glutamate transporters: Confining runaway excitation by shaping synaptic transmission. Nat Rev Neurosci. 2007;8:935–947. doi: 10.1038/nrn2274. [DOI] [PubMed] [Google Scholar]
  • 8.Pasquale EB. Eph-ephrin bidirectional signaling in physiology and disease. Cell. 2008;133:38–52. doi: 10.1016/j.cell.2008年03月01日1. [DOI] [PubMed] [Google Scholar]
  • 9.Klein R. Bidirectional modulation of synaptic functions by Eph/ephrin signaling. Nat Neurosci. 2009;12:15–20. doi: 10.1038/nn.2231. [DOI] [PubMed] [Google Scholar]
  • 10.Murai KK, Nguyen LN, Irie F, Yamaguchi Y, Pasquale EB. Control of hippocampal dendritic spine morphology through ephrin-A3/EphA4 signaling. Nat Neurosci. 2003;6:153–160. doi: 10.1038/nn994. [DOI] [PubMed] [Google Scholar]
  • 11.Jiao JW, Feldheim DA, Chen DF. Ephrins as negative regulators of adult neurogenesis in diverse regions of the central nervous system. Proc Natl Acad Sci USA. 2008;105:8778–8783. doi: 10.1073/pnas.0708861105. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Chaudhry FA, et al. Glutamate transporters in glial plasma membranes: Highly differentiated localizations revealed by quantitative ultrastructural immunocytochemistry. Neuron. 1995;15:711–720. doi: 10.1016/0896-6273(95)90158-2. [DOI] [PubMed] [Google Scholar]
  • 13.Tremblay ME, et al. Localization of EphA4 in axon terminals and dendritic spines of adult rat hippocampus. J Comp Neurol. 2007;501:691–702. doi: 10.1002/cne.21263. [DOI] [PubMed] [Google Scholar]
  • 14.Levenson J, et al. Long-term potentiation and contextual fear conditioning increase neuronal glutamate uptake. Nat Neurosci. 2002;5:155–161. doi: 10.1038/nn791. [DOI] [PubMed] [Google Scholar]
  • 15.LeDoux JE. Emotion circuits in the brain. Annu Rev Neurosci. 2000;23:155–184. doi: 10.1146/annurev.neuro.23.1.155. [DOI] [PubMed] [Google Scholar]
  • 16.Dere E, Huston JP, De Souza Silva MA. The pharmacology, neuroanatomy, and neurogenetics of one-trial object recognition in rodents. Neurosci Biobehav Rev. 2007;31:673–704. doi: 10.1016/j.neubiorev.200701005. [DOI] [PubMed] [Google Scholar]
  • 17.Lehre KP, Rusakov DA. Asymmetry of glia near central synapses favors presynaptically directed glutamate escape. Biophys J. 2002;83:125–134. doi: 10.1016/S0006-3495(02)75154-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Noguchi J, Matsuzaki M, Ellis-Davies GC, Kasai H. Spine-neck geometry determines NMDA receptor-dependent Ca2+ signaling in dendrites. Neuron. 2005;46:609–622. doi: 10.1016/j.neuron.2005年03月01日5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Hayashi Y, Majewska AK. Dendritic spine geometry: Functional implication and regulation. Neuron. 2005;46:529–532. doi: 10.1016/j.neuron.200505006. [DOI] [PubMed] [Google Scholar]
  • 20.Matsuzaki M. Factors critical for the plasticity of dendritic spines and memory storage. Neurosci Res. 2007;57:1–9. doi: 10.1016/j.neures.2006年09月01日7. [DOI] [PubMed] [Google Scholar]
  • 21.Butchbach ME, Tian G, Guo H, Lin CL. Association of excitatory amino acid transporters, especially EAAT2, with cholesterol-rich lipid raft microdomains: Importance for excitatory amino acid transporter localization and function. J Biol Chem. 2004;279:34388–34396. doi: 10.1074/jbc.M403938200. [DOI] [PubMed] [Google Scholar]
  • 22.Zimmer M, Palmer A, Kohler J, Klein R. EphB-ephrinB bi-directional endocytosis terminates adhesion allowing contact mediated repulsion. Nat Cell Biol. 2003;5:869–878. doi: 10.1038/ncb1045. [DOI] [PubMed] [Google Scholar]
  • 23.Sheldon AL, Gonzalez MI, Krizman-Genda EN, Susarla BT, Robinson MB. Ubiquitination-mediated internalization and degradation of the astroglial glutamate transporter, GLT-1. Neurochem Int. 2008;53:296–308. doi: 10.1016/j.neuint.2008年07月01日0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Bach ME, Hawkins RD, Osman M, Kandel ER, Mayford M. Impairment of spatial but not contextual memory in CaMKII mutant mice with a selective loss of hippocampal LTP in the range of the theta frequency. Cell. 1995;81:905–915. doi: 10.1016/0092-8674(95)90010-1. [DOI] [PubMed] [Google Scholar]
  • 25.Gass P, et al. Mice with a fra-1 knockin into the c-fos locus show impaired spatial but regular contextual learning and normal LTP. Brain Res. 2004;130:16–22. doi: 10.1016/j.molbrainres.200407004. [DOI] [PubMed] [Google Scholar]
  • 26.Hunsaker MR, Kesner RP. Dissociations across the dorsal-ventral axis of CA3 and CA1 for encoding and retrieval of contextual and auditory-cued fear. Neurobiol Learn Mem. 2008;89:61–69. doi: 10.1016/j.nlm.2007年08月01日6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Fiala JC, Spacek J, Harris KM. Dendritic spine pathology: Cause or consequence of neurological disorders? Brain Res Brain Res Rev. 2002;39:29–54. doi: 10.1016/s0165-0173(02)00158-3. [DOI] [PubMed] [Google Scholar]
  • 28.Beart PM, O'Shea RD. Transporters for L-glutamate: An update on their molecular pharmacology and pathological involvement. Br J Pharmacol. 2007;150:5–17. doi: 10.1038/sj.bjp.0706949. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Moreno-Flores MT, Wandosell F. Up-regulation of Eph tyrosine kinase receptors after excitotoxic injury in adult hippocampus. Neuroscience. 1999;91:193–201. doi: 10.1016/s0306-4522(98)00568-5. [DOI] [PubMed] [Google Scholar]
  • 30.Rakhade SN, Loeb JA. Focal reduction of neuronal glutamate transporters in human neocortical epilepsy. Epilepsia. 2008;49:226–236. doi: 10.1111/j.1528-1167.2007.01310.x. [DOI] [PubMed] [Google Scholar]
  • 31.Tsuda H, et al. The amyotrophic lateral sclerosis 8 protein VAPB is cleaved, secreted, and acts as a ligand for Eph receptors. Cell. 2008;133:963–977. doi: 10.1016/j.cell.2008年04月03日9. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supporting Information

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences

ACTIONS

RESOURCES

AltStyle によって変換されたページ (->オリジナル) /