Reversal of activity-mediated spine dynamics and learning impairment in a mouse model of Fragile X syndrome

(1)

Article

Reference

Reversal of activity-mediated spine dynamics and learning impairment in a mouse model of Fragile X syndrome

BODA, Bernadett, et al.

Abstract

Fragile X syndrome (FXS) is characterized by intellectual disability and autistic traits, and results from the silencing of the FMR1 gene coding for a protein implicated in the regulation of protein synthesis at synapses. The lack of functional Fragile X mental retardation protein has been proposed to result in an excessive signaling of synaptic metabotropic glutamate receptors, leading to alterations of synapse maturation and plasticity. It remains, however, unclear how mechanisms of activity-dependent spine dynamics are affected in Fmr knockout (Fmr1-KO) mice and whether they can be reversed. Here we used a repetitive imaging approach in hippocampal slice cultures to investigate properties of structural plasticity and their modulation by signaling pathways. We found that basal spine turnover was significantly reduced in Fmr1-KO mice, but markedly enhanced by activity. Additionally, activity-mediated spine stabilization was lost in Fmr1-KO mice. Application of the metabotropic glutamate receptor antagonist α-Methyl-4-carboxyphenylglycine (MCPG) enhanced basal turnover, improved spine stability, but failed to reinstate [...]

BODA, Bernadett, et al . Reversal of activity-mediated spine dynamics and learning impairment in a mouse model of Fragile X syndrome. European Journal of Neuroscience , 2014, vol. 39, no. 7, p. 1130-7

DOI : 10.1111/ejn.12488 PMID : 24712992

Available at:

http://archive-ouverte.unige.ch/unige:41701

Disclaimer: layout of this document may differ from the published version.

1 / 1

(2)

Reversal of activity-mediated spine dynamics and learning impairment in a mouse model of Fragile X syndrome

Bernadett Boda,¹Pablo Mendez,¹Benjamin Boury-Jamot,²Fulvio Magara²andDominique Muller¹

1Department of Basic Neurosciences, School of Medicine, University of Geneva, Geneva 4 1211, Switzerland

2Center for Psychiatric Neurosciences, Cery, Prilly-Lausanne, Switzerland

Keywords: LTD, LTP, memory, spinogenesis, synaptic plasticity, water maze

Abstract

Fragile X syndrome (FXS) is characterized by intellectual disability and autistic traits, and results from the silencing of theFMR1 gene coding for a protein implicated in the regulation of protein synthesis at synapses. The lack of functional Fragile X mental retardation protein has been proposed to result in an excessive signaling of synaptic metabotropic glutamate receptors, leading to alterations of synapse maturation and plasticity. It remains, however, unclear how mechanisms of activity-dependent spine dynamics are affected inFmrknockout (Fmr1-KO) mice and whether they can be reversed. Here we used a repetitive imaging approach in hippocampal slice cultures to investigate properties of structural plasticity and their modulation by signaling pathways.

We found that basal spine turnover was significantly reduced inFmr1-KO mice, but markedly enhanced by activity. Additionally, activity-mediated spine stabilization was lost inFmr1-KO mice. Application of the metabotropic glutamate receptor antagonista- Methyl-4-carboxyphenylglycine (MCPG) enhanced basal turnover, improved spine stability, but failed to reinstate activity-mediated spine stabilization. In contrast, enhancing phosphoinositide-3 kinase (PI3K) signaling, a pathway implicated in various aspects of synaptic plasticity, reversed both basal turnover and activity-mediated spine stabilization. It also restored defective long-term potentiation mechanisms in slices and improved reversal learning inFmr1-KO mice. These results suggest that modulation of PI3K signaling could contribute to improve the cognitive deficits associated with FXS.

Introduction

TheFMR1gene codes for an RNA-binding protein implicated in the shuttling of mRNAs to dendritic sites where it contributes to the local control of protein synthesis. Studies of theFmr1 knockout (Fmr1- KO) mouse model of Fragile X syndrome (FXS) strongly suggest that synaptic defects are an important component of the pathological mechanisms responsible for the disease. Alterations of both the struc- ture and function of excitatory synapses have been reported (Comery et al., 1997; Irwinet al., 2000; Huberet al., 2002; Lauterbornet al., 2007; Huet al., 2008; Auerbach & Bear, 2010). These defects have been proposed to result from an excessive signaling mediated by metabotropic glutamate receptors (Pfeiffer & Huber, 2007) that could affect spine synapse formation and maturation. In vivoexperiments have indeed reported alterations of spine dynamics in the cortex and the presence of unstable spines insensitive to modulation by sensory experience (Cruz-Martinet al., 2010; Panet al., 2010).

Considering that alterations of spine dynamics could represent key mechanisms susceptible to interfere with the development of synaptic networks (Caroniet al., 2012) and thus play a signiﬁcant role in the pathology of FXS, we investigated whether and how these defects could be regulated inFmr1-KO mice. Several different approaches have been proposed or even tested over the last years to counteract the deﬁcient phenotypes observed in Fmr1-KO mice. A main strategy has been to interfere with metabotropic glutamate receptor signaling, which has been shown to correct many behav-

ioral phenotypes in Fmr1-KO mice (Dolen et al., 2007; Michalon et al., 2012; Osterweil et al., 2012). There are, however, also studies suggesting that it is possible to reverse selective deﬁcits in Fmr1-KO mice by acting on other signaling pathways, such as protein synthesis regulation, ERK1/2, PAK or brain-derived neurotrophic factor (BDNF) signaling (Hayashiet al., 2007; Lauterbornet al., 2007; Huet al., 2008; Chenet al., 2010; Bhattacharyaet al., 2012;

Dolanet al., 2013; Osterweil et al., 2013). These results thus suggest the involvement of multiple signaling pathways in the pheno- typic alterations observed in FXS. Consistent with this, Fmr1-KO mice not only show an enhanced long-term depression (LTD), but also defects in long-term potentiation (LTP), although this issue has been debated and mainly reported at neocortical synapses (Lauter- bornet al., 2007; Meredithet al., 2007; Huet al., 2008; Krueger &

Bear, 2011). Defects in LTP could be rescued by application of BDNF or by enhancing phosphoinositide-3 kinase (PI3K) signaling (Lauterborn et al., 2007; Huet al., 2008). The role of this pathway remains, however, controversial, as it was found to be either com- promised or upregulated in different studies (Huet al., 2008; Gross et al., 2010; Sharmaet al., 2010).

Here we focused speciﬁcally on activity-dependent spine dynamics in the hippocampus, and investigated whether the alterations observed in Fmr1-KO mice could be reversed bya-Methyl-4-carboxyphenylglycine (MCPG), a selective group I/II metabotropic glutamate receptor antagonist, and by dipotassium bisperoxo(5-hydroxypyridine-2- carboxyl)oxovanadate (BpV), a phosphatase and tensin homolog (PTEN) inhibitor that enhances PI3K-protein kinase B (AKT) signaling. Our results indicate that the PTEN inhibitor, but not MCPG, is able to restore spine dynamics alterations inFmr1-KO mice and to improve cognitive functions in a behavioral task in these mice.

Correspondence: Dr D. Muller, as above.

E-mail: [email protected]

Received 22 October 2013, revised 17 December 2013, accepted 18 December 2013

European Journal of Neuroscience, Vol. 39, pp. 1130–1137, 2014 doi:10.1111/ejn.12488

(3)

Materials and methods

Organotypic cultures and transfection

Hippocampal slice cultures (400lm) were prepared from Fmr1-KO mice (Mientjeset al., 2006) and wild-type (WT) littermates at post- natal day 5, and cultured for 3 weeks (Stoppini et al., 1991). The experimental protocols were reviewed and approved by the Ethics Committee of the University Medical Center of Geneva, by the Can- tonal Veterinary Ofﬁce, Geneva, Switzerland, and were carried out in conformity with the guidelines laid down by the NIH in the USA regarding the care and use of animals for experimental procedures.

Transfection was performed with the gene gun method (Helios Gene Gun; Bio-Rad) at 8–10 days in vitro using different plasmid con- structs (pcDNA3.1-EGFP and pCX-mRFP1; ratio of plasmid: 10lg for 5 mg of 1.6lm gold microcarriers). The shRNA construct was kindly provided by I. Caille (Universite Pierre et Marie Curie, Paris, France), and its efﬁcacy demonstrated both in in vitro and in vivo experiments (Scotto-Lomassese et al., 2011). A negative control small interference sequence was obtained from Qiagen Allstars.

(S)-MCPG (Tocris Bioscience) was used at 100 l^M. BpV (Enzo Life Sciences and Calbiochem/Millipore) was used at 15 nM in culture medium, and 30lg/100 g for intraperitoneal (i.p.) injections.

Confocal imaging and electrophysiology

Four days after transfection, slice cultures were imaged repetitively using an Olympus Fluoview confocal microscope as described (Dubos et al., 2012). Secondary or tertiary dendritic segments (35–45lm) on apical dendrites of CA1 pyramidal neurons in the stratum radiatum were visualized at 0, 5, 24, 48 and 72 h. Z-stack images were analysed using Osirix software and cross-checked with another experimenter. All protrusions were included in the analysis, and stability was assessed as the fraction of spines still present at later times. Spines were considered to enlarge if the spine head width increased by more than 0.1 lm between the 0 and 5 h time points. All data are given withnbeing the number of dendritic segments analysed considering one dendritic segment per neuron and one neuron per slice culture. When appropriate, we also give the total number of spines analysed in the pool of experiments.

For electrophysiological recordings, hippocampal cultures or acute slices (4–6-week-old mice) were maintained in an interface chamber as described (De Rooet al., 2008b), perfused in a medium with the following composition (in mM): NaCl, 124; KCl, 1.6; MgCl, 1.5; CaCl₂, 2.5; KH2PO4, 1.2; NaHCO3, 24; glucose, 10; ascorbic acid, 2; bubbled with 95% O2and 5% CO2. Theta-burst stimulation (TBS) consisted of ﬁve trains at 5 Hz, each composed of four pulses at 100 Hz, repeated three times at 10-s intervals. Field excitatory postsynaptic potential slopes and amplitudes were measured using IGOR software.

Western blot

The effect of BpV on AKT phosphorylation on the S473 site was tested in hippocampal slice cultures maintained 12–14 daysin vitro and treated with 15 n^M BpV, as well as on 4–6-week-old mice injected i.p. with BpV [30lg/100 g in saline (200lL)]. The tissue was lysed on ice 30 min or 1–4 h after treatments for slices and animals, respectively, using a solution containing (in mM): NaCl, 150;

Tris-HCl, 50, pH 7.4; EDTA, 2; dithiothreitol, 1; 1% Triton X-100, 10% glycerol, complete protease inhibitor tablet (Roche), completed with phosphatase inhibitors (in mM): sodium orthovanadate, 1; para- nitrophenylphosphate, 20; b-glycerophosphate, 20; okadaic acid,

50 ng/mL. Lysates were sonicated, and 30lg of total protein put on Nupage 4–12% Bis-Tris gel (Invitrogen). Gels were run in 3-(N- morpholino)propanesulfonic acid sodium dodecyl sulfate buffer (Invitrogen) and transferred to nitrocellulose membranes. Phospho- AKT immunoblotting was performed using 1 : 3000 dilution of the primary rabbit monoclonal antibody against phosphoSer473 (Cell Signaling Technology). Densitometric analysis of the pAKT band of 60 kDa was normalized by b-actin immunoblot and expressed as a function of WT values (monoclonal, 1 : 20 000; Sigma-Aldrich).

Animals and Morris water maze

Twenty male adult Fmr1-KO mice (FVB.129P2-Pde6b⁺ Tyr^c-ch Fmr1^tm1Cgr/J) and 20 male adult WT controls (FVB.129P2-Pde6b⁺ Tyr^c-ch/AntJ) were purchased from Jackson Laboratories. Mice were individually identiﬁed and housed according to the Swiss legislation on animal protection. OneFmr1-KO and one WT mice were used for pilot tests; theﬁnal sample (19Fmr1-KO and 19 WT) was divided into two groups: nine mice were treated with BpV and 10 received the vehicle. The mice were trained in a Morris water maze with four daily trials during four consecutive days. At day 5, a probe trial was run in the absence of the escape platform. Immediately after, mice received either an i.p. injection of BpV [30lg/100 g in saline (200lL)] or an i.p. injection of 200lL saline. One hour later, all mice started a novel, four-trial training session with the platform in the opposite quadrant (reversal learning). Reversal training continued for two more days; treatment with BpV or saline was repeated accordingly. At day 8, a second probe test was run. Swim paths were videorecorded and analysed by a videotracking system (EthoVision, Noldus, the Netherlands). Variables assessed were escape latencies, times spent in target quadrants and platform place proximity indices (PIs), calculated by measuring the average distances to the platform [(PI probe1 PI probe2)/(PI probe1+PI probe2)9100].

Statistics

Data are represented as meanSEM. Statistical analyses were carried out using Student’st-tests unless otherwise mentioned.

Results

Decreased spine turnover inFmr1-deficient hippocampal neurons

Spine dynamics was analysed using a repetitive imaging approach applied to transfected CA1 hippocampal neurons in slice cultures prepared fromFmr1-KO mice or WT littermates. Under basal conditions in WT slice cultures, a substantial fraction of spines are gener- ated and eliminated every day (De Rooet al., 2008a). InFmr1-KO mice slices, this basal spine turnover was signiﬁcantly reduced, and affected both newly formed and lost spines (Fig. 1A–E; new spines: 9.81.2%, n=10, 452 spines, Fmr1-KO vs.

16.5 0.8%, n=11, 348 spines, WT, P<0.001; lost spines:

9.51.5%, Fmr1-KO vs. 16.22%, WT, P <0.05). To further conﬁrm this result, we then also transfected CA1 neurons with EGFP and either a small hairpin construct directed againstFmr1(shFmr1) or with a control, non-silencing sequence (shCtrl). The efﬁcacy of the shFmr1 construct has been demonstrated to reduce Fragile X mental retardation protein expression levels both underin vitroand in vivo conditions (Scotto-Lomassese et al., 2011). Acute transfection of slice cultures with this shFmr1 construct reproduced the changes observed inFmr1-KO mice, and reduced both the fraction

Reversal of spine dynamics in Fragile X syndrome 1131

(4)

of newly formed spines per 24 h (10.4 1.9%,n=9, 204 spines, ShFmr1 vs. 17.61.7%, n=7, 219 spines, shCtrl, P<0.05) and the fraction of eliminated spines (13.01.5%, shFmr1 vs.

21.04.3%, shCtrl, P=0.07). No differences in spine density could be detected between these conditions (Fig. 1F). Finally, to further characterize these deficits, we also carried out a morphological analysis of spine size and spine length in Fmr1-KO and shFmr1- transfected cells. The results illustrated in Fig. 1G revealed a significant increase in mean spine length in both Fmr1-KO and shFmr1-transfected neurons, associated with a significant decrease in mean spine width (Fig. 1G; n=452 and 348 spines analysed;

**P<0.01,***P<0.001). Note that the mean spine length in CA1 pyramidal neurons transfected with the scrambled, control shRNA was also signiﬁcantly shorter than in non-transfected cells (P<0.001). We do not know the reasons for these differences, but previous work has shown that RNA interference approaches could have off-target non-speciﬁc effects on dendritic spines (Alvarez et al., 2006). These results suggest the existence of an increased number of immature spines inFmr1-KO that is consistent with previous observations (Comeryet al., 1997; Grossmanet al., 2006).

Defects in activity-dependent spine dynamics in hippocampal neurons ofFmr1-KO mice

We then assessed the effects of activity on spine turnover. Applica- tion of TBS, a protocol that induces LTP in slices cultures, resulted

in a marked increase in spine formation and elimination over the next 24 h (Fig. 2A, C and D; new spines: 42.47.9%,n=5, 240 spines analysed, P<0.01; lost spines: 43.7 8.8%, n=5, P<0.01), consistent with previous observations (De Roo et al., 2008b). Additionally, analyses of spine stability and comparisons between spines that did and did not enlarge 5 h after TBS (Fig. 2E) showed that enlarged spines were also more likely to persist over the next 48 h than spines that did not enlarge (Fig. 2F;

87.5 12.5%, n=4 vs. 55.55.0%, n=4, P<0.01). This is consistent with the notion that application of TBS and induction of plasticity can lead to a selective stabilization of stimulated synapses (De Roo et al., 2008b). In Fmr1-KO mice, spine turnover was reduced under basal conditions (Fig. 2B), but showed a fourfold increase following TBS, reaching similar levels as those observed in WT mice (Fig. 4C and D: new spines: 426.4%, n=5, 199 spines, P<0.001; lost spines: 37.85.5%, n=5, P<0.001).

Thus, the sensitivity of spine turnover to activity was strongly enhanced inFmr1-KO slices. Furthermore, the selective stabilization of the spines that enlarged following TBS was lost in Fmr1-KO slices: both spines that did and did not enlarge showed the same low 48 h stability (Fig. 2F; 56.515.3%, n=4 vs. 53.06.1%, n=4). The proportion of spine that enlarged was, however, not sig-

A B C

D E

F G

Fig. 1.Alteration of spine dynamics in theFmr1-KO mouse model of Frag- ile X syndrome (FXS). (A–C) Illustrations of 24-h spine turnover in transfected pyramidal neurons of (A) WT and (B)Fmr1-KO slice cultures, and of a WT pyramidal neuron transfected with shFmr1 (C; bar: 2lm). (D) Decrease in basal spine growth mechanisms detected over 24 h inFmr1-KO and shFmr1-transfected neurons (n=7–11). (E) Same, but for disappearing spines. (F) Absence of changes in spine density observed under the same conditions. (G) Increase in spine length and decrease in spine head width observed in Fmr1-KO and shFmr1-transfected cells (n=452, 348;

*P<0.05,**P<0.01,***P<0.001).

A B

C D

E F

F G

Fig. 2. Alteration of activity-dependent spine dynamics inFmr1-KO mice.

(A and B) Illustration of the changes in spine dynamics induced by theta- burst stimulation (TBS) in slice cultures of WT and Fmr1-KO mice (bar:

2lm). (C) Increase in spine growth detected over 24 h and induced by TBS in slice cultures of WT andFmr1-KO mice (n=5). (D) Same, but for lost spines. (E) Illustration of the 48-h stability of spines that enlarged as a result of TBS in slice cultures from WT (left) and Fmr1-KO (right) mice (bar:

2lm). (F) Stability over 48 h of spines that enlarged and did not enlarge as a result of TBS in WT and Fmr1-KO mice (n=5; two-way ANOVA;

*P<0.05,**P<0.01,***P<0.001).

(5)

niﬁcantly different between Fmr1-KO and WT slices (21.82.8%

vs. 16.41.9%). Thus,Fmr1-KO mice showed an increased sensitivity of spine dynamics to activity and a loss of activity-mediated spine stabilization.

Partial reversal of spine dynamics deficits inFmr1-KO mice by a metabotropic glutamate receptor antagonist

We then investigated whether these defects could be reversed, and tested the effects of the metabotropic glutamate receptor antagonist MCPG (100 l^M) on spine dynamics. In WT slice cultures, MCPG had no detectable effects on basal spine formation or elimination mechanisms (Fig. 3A and B). However, it prevented the increase in spine turnover triggered by TBS (Fig. 3A and B; new spines:

23.83.3%, n=6, 233 spines vs. 42.4 7.9%, n=5, 240 spines, P <0.05; lost spines: 20.8 2.4% vs. 43.7 8.8%, P<0.05). It also interfered with the mechanisms of activity-dependent spine stabilization. In WT mice, MCPG signiﬁcantly reduced the differential stability of enlarged vs. non-enlarged spines following TBS (Fig. 3C; circles: ns: non-signiﬁcant difference, two-way

ANOVA), consistent with evidence suggesting an involvement of metabotropic receptors in LTP induction mechanisms (Anwyl, 2009). In Fmr1-KO mice, MCPG had three main effects. First, it restored the level of basal turnover by enhancing both newly formed and lost spines per 24 h (Fig. 3A and B: new spines: 17.03.2%

vs. 9.31.3%, n=6, 8, P<0.05; lost spines: 15.22.5% vs.

9.51.5%, n=6, 8, P <0.05), consistent with the idea that the decreased turnover observed in Fmr1-KO mice could be due to an over-activation of metabotropic receptors (Pfeiffer & Huber, 2007).

Second, it reduced the increased sensitivity of spine turnover to activity: new spine formation still increased following TBS, but to a lesser extent (Fig. 3A; 27.23.6%, n=5 cells, 237 spines vs.

15.6 3.0%, n=7, 274 spines, P<0.05). However, the changes in spine loss following TBS stimulation did not reach statistical sig- niﬁcance (Fig. 3B;P=0.20). Thus, MCPG normalized basal spine dynamics and reduced the sensitivity of spine turnover to activity.

Third, MCPG improved the general stability of spines [Fig. 3D;

compare with (circles) and without MCPG (squares),P<0.05, two- way^ANOVA]. However, MCPG failed to improve the defects in activity-dependent spine stabilization. Enlarged and non-enlarged spines showed the same stability following TBS, and there was no prefer- ential stabilization of enlarged spines (Fig. 3D; circles, ns: non-sig- niﬁcant, n=5 cells, 237 spines analysed). Thus, MCPG improved the general stability of spines but did not restore mechanisms of activity-dependent spine stabilization.

Reversal of spine dynamics deficits inFmr1-KO mice by PTEN inhibition

As the reduced turnover and loss of activity-dependent stability observed inFmr1-KO mice suggested a deficit in LTP-mediated signaling mechanisms, we then investigated whether enhancing PI3K signaling could also reverse these alterations as previously shown for other LTP-associated deficits (Lauterbornet al., 2007; Huet al., 2008). To do this, we used BpV, a PTEN inhibitor that is selective when used at low nanomolar concentrations, and has been shown to increase PI3K activity and AKT phosphorylation even underin vivo conditions (Schmid et al., 2004; Jurado et al., 2010; Mao et al., 2013). Wefirst verified that 15 nMof BpV applied to slice cultures or 30lg/100 g given intraperitoneally underin vivoconditions was able to increase AKT phosphorylation, assessed using a phospho- specific antibody against the S473 site of AKT. Figure 4A illustrates the increase in AKT phosphorylation of S473 obtained 1 h after injection of BpV (30lg/100 g) intraperitoneally in WT andFmr1- KO mice. The experiments were also reproduced in slice cultures, and S473 phosphorylation assessed 30 min after addition of BpV to the culture medium. All quantitative data normalized to the optical densities measured in WT tissue are shown in Fig. 4B (n=3–9).

The results indicate that BpV at the concentrations used promoted phosphorylation of AKT on its S473 site both under in vitro and in vivo conditions, and in WT and Fmr1-KO mice. Note that our data also indicate a reduced phosphorylation of the S473 site of AKT inFmr1-KO mice as compared with WT animals.

We then examined the effects of BpV on spine turnover in WT and Fmr1-KO mice. In WT slice cultures, BpV signiﬁcantly increased the level of basal turnover, enhancing spine formation mechanisms to the level observed following TBS, occluding further effects of TBS (Fig. 5A; BpV: 41.84.6%, n=5, 168 spines;

TBS: 42.47.9%, n=5, 133 spines). BpV also increased spine elimination in a similar manner (Fig. 5B). BpV, however, did not alter mechanisms of activity-mediated spine stabilization, and the differential survival of enlarged and non-enlarged spines observed after TBS was not modiﬁed (Fig. 5C; stable spines at 48 h:

77.6 8.0%, n=5 vs. 60.06.5%, n=5, P<0.05; two-way

ANOVA).

In Fmr1-KO mice, BpV reversed all the defects of spine dynamics. BpV increased basal turnover both regarding new spine formation and spine elimination (Fig. 5A; KO, new spines: 9.21.3%, n=10, 348 spines; KO+BpV, new spines: 22.8 2.7%, n=6, 277 spines, P<0.01; Fig. 5B; KO, lost spines: 9.5 1.5%;

KO+BpV: 19.4 2.8%, P<0.01). Furthermore, a sensitivity to activity was preserved and the level of turnover could still be increased by TBS application (Fig. 5A and B; KO+BpV+TBS, new spines: 33.82.2%, n=11, 369 spines, P<0.01;

A B

C D

Fig. 3. Effects of MCPG on spine dynamics and activity-dependent stabilization in WT andFmr1-KO mice. (A) Changes in spine formation induced by MCPG under basal conditions (gray bars) and following application of TBS (black bars) in WT andFmr1-KO mice. (B) Same, but for lost spines.

(C) Stability over 48 h of enlarged and non-enlarged spines after TBS in MCPG-treated and non-treated WT mice (n=5). Note that MCPG prevented the differential spine stabilization in WT slices. (D) Same but forFmr1-KO mice (n=5). Note that MCPG did not improve the loss of effect observed inFmr1-KO mice (see Fig. 2).

(6)

KO+BpV+TBS, lost spines: 29.0 3.0%, P<0.05). Most important, BpV also reversed the loss of activity-mediated spine stabilization observed in Fmr1-KO mice: enlarged spines were now differentially stabilized when compared with non-enlarged spines (Fig. 5D; 82.45.1% vs. 62.8 5.4%,n=9,P <0.01, two-way

ANOVA, Bonferronipost hoctest).

Reversal of LTP and learning deficits inFmr1-KO mice We then wondered whether BpV treatment could also have functional implications. For this, we first tested the effects of BpV on LTP induction mechanisms on acute hippocampal slices of young adult mice (4–6-week-old mice). As shown in Fig. 6A, and consistent with other studies (Lauterborn et al., 2007; Meredith et al., 2007; Huet al., 2008), we found that LTP was decreased in Fmr1- KO mice (Fig. 6A and B; 15.5 3.0% vs. 37.15.8%, n=6–8, P<0.01). Application of BpV (15 nM) had no detectable effects on basal transmission, but partially reversed the decrease in LTP measured at 30 min (28.90.9% vs. 15.5 3.0%, n=6–8, P<0.01), without significantly affecting LTP in WT slices. This result thus suggested that BpV could be beneficial for LTP-dependent learning in Fmr1-KO mice. We thus tested whether and how BpV affected adultFmr1-KO mice behavior in a Morris water maze task. Four groups of mice, WT (n=19) and Fmr1-KO mice (n=19), with (n=9) and without (n=10) treatment of BpV were compared. Mice underwent 4 days of training and were then submitted to a probe trial without platform. After this, they were retested for reversal learning over three consecutive days with BpV [30lg/

100 g in saline (200lL)] or saline (200lL) injected before each test. Finally they were submitted to a final probe trial. Before any treatment, Fmr1-KO and WT mice learned equally well to localize the platform position (Fig. 6C). In contrast, Fmr1-KO mice were significantly less proficient than WT in the reversal task (Fig. 6, open circles in D and E; P<0.001, repeated-measures ANOVA).

Comparison of BpV-treated and vehicle-injected WT mice showed no detectable differences. However, BpV signiﬁcantly improved the learning curve ofFmr1-KO mice in the reversal task (Fdf 1,17=2.9, P<0.05, repeated-measures ANOVA; Fig. 6E). BpV treatment also improved the new target quadrant preference of Fmr1-KO mice in the second probe trial (KO+BpV: 21.9 2.7 s; WT:

20.3 2.0 s; KO: 17.6 2.2 s, P<0.01, ^ANOVA). Furthermore, analysis of the PI reﬂecting the average distance to the target platform in the second probe trial showed that the searching precision improved slightly for both WT mice and BpV-treated Fmr1-KO mice, while vehicle-treated Fmr1-KO mice clearly worsened their performance (Fig. 6F; P<0.05). Together these results provide strong support to the rescuing potential of BpV in this learning task.

Discussion

Multiple lines of evidence indicate that FXS results from a dysregulation of mechanisms of protein synthesis at synapses that could be linked to an exaggerated signaling through group I metabotropic glutamate receptors (Osterweil et al., 2012). Accordingly, the main stream strategy currently used to try to reverse the deﬁcits associated with the disease includes the development of selective antagonists of these receptors (Dolen et al., 2007; Krueger & Bear, 2011; Mich- alon et al., 2012). A few other studies, however, also suggest that interfering with other signaling pathways could be beneﬁcial for reversing behavioral phenotypes in Fmr1-KO mice (Hayashi et al., 2007; Lauterborn et al., 2007; Dolan et al., 2013; Osterweil et al., 2013). Here we focused on spine dynamics alterations in the hippo-

A B

C D

Fig. 5.Effects of BpV on spine dynamics and activity-dependent stabilization in WT andFmr1-KO mice. (A) Changes in spine formation induced by BpV under basal conditions (gray bars) and following application of TBS (black bars) in WT andFmr1-KO mice (n=4–10;*P<0.05,**P<0.01,

***P<0.001). (B) Same, but for lost spines. (C) Stability over 48 h of enlarged and non-enlarged spines after TBS in BpV-treated and non-treated WT slices (n=5;*P<0.05, two-wayANOVA). Note that BpV treatment did not affect the differential stabilization in WT slices. (D) Same but forFmr1- KO mice (n=5; **P< 0.01, two-wayANOVA). Note that BpV rescued the defect inFmr1-KO slices (see Fig. 2).

A

B

Fig. 4.Enhancement of AKT phosphorylation on the S473 site by BpV.

(A) Western blot illustrating the increased phosphorylation of AKT produced by BpV (15 nM) in slice cultures (left) and inin vivo Fmr1-KO mice (right).

(B) Quantitative assessment of AKT phosphorylation on the S473 site produced by BpV in slice cultures (in vitro: 15 nM,n=3) and following intraperitoneal injection in mice (in vivo: 30lg/100 g, n=3–5; *P<0.05,

**P<0.01).

(7)

campus, and provide evidence that promoting PI3K signaling reverses the spine dynamics deﬁcits and improves LTP and cognitive functions in this model.

First, our study extends to the hippocampus previous in vivo analyses indicating that properties of structural plasticity are affected in the somatosensory cortex of Fmr1-KO mice (Cruz-Martin et al., 2010; Pan et al., 2010). In the hippocampus, weﬁnd that the level of turnover of spine synapses is signiﬁcantly reduced under basal conditions, with a decreased formation and elimination of spines.

The effect was observed both inFmr1-KO mice and following acute knockdown of the Fragile X mental retardation protein using a shRNA approach. Our results, however, do not suggest an increase in spine elimination as proposed by some studies (Pfeiffer et al., 2010). On the other hand, we find that spine dynamics show an increased sensitivity to activity, and furthermore that the mechanisms of spine stabilization induced by activity are deficient. This result is consistent with recent observations indicating that signaling pathways and molecules important for LTP stability are defective in Fmr1-KO mice (Chenet al., 2010; Seeseet al., 2012). Thus, activity strongly stimulates the formation of unstable synapses, a result that is in line with the conclusion of anin vivostudy carried out in the somatosensory cortex of adult mice (Pan et al., 2010). In this and another in vivostudy, however, the level of basal turnover was enhanced and not reduced (Cruz-Martinet al., 2010). Spine turnover was, however, poorly sensitive to sensory experience, suggesting that it was already maximally activated. It might thus be that the discrepancies observed in basal level of spine turnover could reflect differences in the levels of neuronal activity. This might also explain why under dissociated culture conditions, where spontaneous activity is usually high, spine elimination could be enhanced (Pfeiffer et al., 2010). All together, these different studies suggest that an important component of the synaptic alterations present inFmr1-KO

mice is a deﬁcit in the mechanisms of activity-dependent synapse stabilization (Chenet al., 2010; Seeseet al., 2012), resulting in the formation of an excessive number of unstable spines. A similar conclusion was reached in analyses of PAK3 KO mice, another mouse model of intellectual disability, where spine dynamics is also signiﬁ- cantly altered and associated with an increased formation of unstable spines (Duboset al., 2012).

A second interesting result of this study is the possibility to differentially reverse these spine dynamics deficits by MCPG, a group 1 metabotropic glutamate receptor antagonist, and by BpV, a PTEN inhibitor that enhances PI3K signaling (Schmidet al., 2004; Jurado et al., 2010; Mao et al., 2013). MCPG restored a control level of basal turnover, it also improved global spine stability in Fmr1-KO mice, but did not reverse the deficit in activity-mediated spine stabilization. This suggests a differential contribution of metabotropic glutamate receptors to spine dynamics and spine stability that could be mediated by different signaling pathways. Notably the activity- dependent spine stabilization mechanisms probably require to maintain some functionality of metabotropic glutamate receptor signaling, consistent with the notion that MCPG interferes with induction of LTP (Anwyl, 2009). This is also in line with the notion that activity-dependent spine stabilization may depend upon specific signaling systems and regulations of the cytoskeleton that are defective in Fmr1-KO mice (Chen et al., 2010; Seese et al., 2012). However, although MCPG did not fully reverse the spine dynamics deficit reported here, it is interesting that chronic block- ade of mGluR5 was recently reported to restore many cortical deficits inFmr1-KO mice, including altered dendritic spine phenotypes, excessive LTD, mammalian target of rapamycin signaling and avoidance learning (Michalonet al., 2012).

In addition to MCPG, our study suggests that another interesting molecule to reverse phenotypes inFmr1-KO mice is BpV, a PTEN

A B

C D E F

Fig. 6. BpV enhances LTP in slices ofFmr1-KO mice and improves behavior in a reversal learning task. (A) Illustration of the LTP obtained in slices of WT andFmr1-KO mice with and without treatment with BpV (n=6–8). (B) Increase in LTP observed 30 min after stimulation obtained under the same conditions (n=6–8;*P<0.05;**P<0.01). (C) Escape latencies of WT andFmr1-KO mice in a water maze place learning task (n=19 per group). (D) Escape latencies of BpV-treated and non-treated WT mice (n=9–10) in the reversal task. (E) Same, but for BpV-treated and non-treated Fmr1-KO mice (n=9–10;

*P<0.05,ANOVA). (F) Proximity index (PI) to the novel (i.e. reversed) as compared with the old platform position for the four groups of mice (*P<0.05,

**P<0.01).

(8)

inhibitor. Application of BpV enhances PI3K activity, a signaling pathway activated by many surface receptors including BDNF that also inhibits LTD (Juradoet al., 2010) and contributes to LTP signaling (Tanget al., 2002). Consistent with this, our experiments show that BpV could restore activity-dependent spine stabilization. These results thus suggest that in addition to an exaggerated LTD type of signaling, another defect present in Fmr1-KO mice could involve defective LTP mechanisms. Our LTP experiments support this inter- pretation and are in line with those of several other studies (Lauter- bornet al., 2007; Meredithet al., 2007; Huet al., 2008; but see also Krueger & Bear, 2011). A question, however, remains regarding the specific role of PI3K signaling in FXS. Two recent studies showed that PI3K activity and AKT phosphorylation are already upregulated inFmr1-KO mice (Grosset al., 2010; Sharmaet al., 2010), a result at variance with another study (Huet al., 2008). It should be noted, however, that the PI3K-AKT signaling pathway is downstream of many cell surface receptors and implicated in many interactions with other systems, possibly through differential regulations of AKT phosphorylation sites (Hemmings & Restuccia, 2012). Although a precise understanding of the complex signaling mechanisms implicated in PI3K signaling is beyond the scope of this study, our data clearly show that BpV treatment enhanced AKT phosphorylation on the S473 site both underin vitroandin vivoconditions, and was able not only to reverse the deficits in spine dynamics observed inFmr1-KO mice, but also to increase LTP and improveFmr1-KO mice behavior in a Morris water maze learning task. Consistent with these results, PI3K signaling has been shown to contribute to LTP mechanisms (Tanget al., 2002), to be involved in BDNF signaling, which also rescued LTP inFmr1-KO mice (Lauterbornet al., 2007), to regulate synaptogenesis and spinogenesis in hippocampal neurons (Jaworski et al., 2005; Cuestoet al., 2011), and more importantly to be also deficient in a mouse model of Angelman syndrome, another neurode- velopment disorder with severe cognitive deficits and autistic traits (Caoet al., 2013). It might therefore represent an interesting target for reversing cognitive deficits.

Conflict of interest

The authors declare no competingﬁnancial interests.

Acknowledgements

The authors are grateful to Mrs Lorena Jourdain for excellent technical help.

This work was supported by the Swiss National Science Foundation (grant 310030B_144080/1) and by the NCCR Synapsy.

Abbreviations

AKT, protein kinase B; BDNF, brain-derived neurotrophic factor; BpV, dipotassium bisperoxo(5-hydroxypyridine-2-carboxyl)oxovanadate; FXS, Fragile X syndrome; KO, knockout; LTD, long-term depression; LTP, long-term potentiation; MCPG,a-Methyl-4-carboxyphenylglycine; PI, proximity index;

PI3K, phosphoinositide-3 kinase; PTEN, phosphatase and tensin homolog;

TBS, theta-burst stimulation; WT, wild-type.

References

Alvarez, V.A., Ridenour, D.A. & Sabatini, B.L. (2006) Retraction of synapses and dendritic spines induced by off-target effects of RNA interference.J. Neurosci.,26, 7820–7825.

Anwyl, R. (2009) Metabotropic glutamate receptor-dependent long-term potentiation.Neuropharmacology,56, 735–740.

Auerbach, B.D. & Bear, M.F. (2010) Loss of the fragile X mental retardation protein decouples metabotropic glutamate receptor dependent priming of

long-term potentiation from protein synthesis. J. Neurophysiol., 104, 1047–1051.

Bhattacharya, A., Kaphzan, H., Alvarez-Dieppa, A.C., Murphy, J.P., Pierre, P. & Klann, E. (2012) Genetic removal of p70 S6 kinase 1 corrects molec- ular, synaptic, and behavioral phenotypes in fragile X syndrome mice.

Neuron,76, 325–337.

Cao, C., Rioult-Pedotti, M.S., Migani, P., Yu, C.J., Tiwari, R., Parang, K., Spaller, M.R., Goebel, D.J. & Marshall, J. (2013) Impairment of TrkB- PSD-95 signaling in Angelman syndrome.PLoS Biol.,11, e1001478.

Caroni, P., Donato, F. & Muller, D. (2012) Structural plasticity upon learning: regulation and functions.Nat. Rev. Neurosci.,13, 478–490.

Chen, L.Y., Rex, C.S., Babayan, A.H., Kramar, E.A., Lynch, G., Gall, C.M.

& Lauterborn, J.C. (2010) Physiological activation of synaptic Rac>PAK (p-21 activated kinase) signaling is defective in a mouse model of fragile X syndrome.J. Neurosci.,30, 10977–10984.

Comery, T.A., Harris, J.B., Willems, P.J., Oostra, B.A., Irwin, S.A., Weiler, I.J. & Greenough, W.T. (1997) Abnormal dendritic spines in fragile X knockout mice: maturation and pruning deﬁcits. Proc. Natl. Acad. Sci.

USA,94, 5401–5404.

Cruz-Martin, A., Crespo, M. & Portera-Cailliau, C. (2010) Delayed stabilization of dendritic spines in fragile X mice.J. Neurosci.,30, 7793–7803.

Cuesto, G., Enriquez-Barreto, L., Carames, C., Cantarero, M., Gasull, X., Sandi, C., Ferrus, A., Acebes, A. & Morales, M. (2011) Phosphoinositide- 3-kinase activation controls synaptogenesis and spinogenesis in hippocampal neurons.J. Neurosci.,31, 2721–2733.

De Roo, M., Klauser, P., Mendez, P., Poglia, L. & Muller, D. (2008a) Activ- ity-dependent PSD formation and stabilization of newly formed spines in hippocampal slice cultures.Cereb. Cortex,18, 151–161.

De Roo, M., Klauser, P. & Muller, D. (2008b) LTP promotes a selective long-term stabilization and clustering of dendritic spines. PLoS Biol., 6, e219.

Dolan, B.M., Duron, S.G., Campbell, D.A., Vollrath, B., Rao, B.S., Ko, H.Y., Lin, G.G., Govindarajan, A., Choi, S.Y. & Tonegawa, S. (2013) Rescue of fragile X syndrome phenotypes in Fmr1 KO mice by the small-molecule PAK inhibitor FRAX486. Proc. Natl. Acad. Sci. USA, 110, 5671–5676.

Dolen, G., Osterweil, E., Rao, B.S., Smith, G.B., Auerbach, B.D., Chattarji, S. & Bear, M.F. (2007) Correction of fragile X syndrome in mice.Neuron, 56, 955–962.

Dubos, A., Combeau, G., Bernardinelli, Y., Barnier, J.V., Hartley, O., Gaert- ner, H., Boda, B. & Muller, D. (2012) Alteration of synaptic network dynamics by the intellectual disability protein PAK3. J. Neurosci., 32, 519–527.

Gross, C., Nakamoto, M., Yao, X., Chan, C.B., Yim, S.Y., Ye, K., Warren, S.T. & Bassell, G.J. (2010) Excess phosphoinositide 3-kinase subunit synthesis and activity as a novel therapeutic target in fragile X syndrome.J.

Neurosci.,30, 10624–10638.

Grossman, A.W., Elisseou, N.M., McKinney, B.C. & Greenough, W.T.

(2006) Hippocampal pyramidal cells in adult Fmr1 knockout mice exhibit an immature-appearing proﬁle of dendritic spines.Brain Res.,1084, 158– 164.

Hayashi, M.L., Rao, B.S., Seo, J.S., Choi, H.S., Dolan, B.M., Choi, S.Y., Chattarji, S. & Tonegawa, S. (2007) Inhibition of p21-activated kinase rescues symptoms of fragile X syndrome in mice. Proc. Natl. Acad. Sci.

USA,104, 11489–11494.

Hemmings, B.A. & Restuccia, D.F. (2012) PI3K-PKB/Akt pathway. Cold Spring Harb. Perspect. Biol.,4, a011189.

Hu, H., Qin, Y., Bochorishvili, G., Zhu, Y., van Aelst, L. & Zhu, J.J. (2008) Ras signaling mechanisms underlying impaired GluR1-dependent plasticity associated with fragile X syndrome.J. Neurosci.,28, 7847–7862.

Huber, K.M., Gallagher, S.M., Warren, S.T. & Bear, M.F. (2002) Altered synaptic plasticity in a mouse model of fragile X mental retardation.Proc.

Natl. Acad. Sci. USA,99, 7746–7750.

Irwin, S.A., Galvez, R. & Greenough, W.T. (2000) Dendritic spine structural anomalies in fragile-X mental retardation syndrome. Cereb. Cortex, 10, 1038–1044.

Jaworski, J., Spangler, S., Seeburg, D.P., Hoogenraad, C.C. & Sheng, M.

(2005) Control of dendritic arborization by the phosphoinositide-3′-kinase- Akt-mammalian target of rapamycin pathway. J. Neurosci., 25, 11300–

11312.

Jurado, S., Benoist, M., Lario, A., Knafo, S., Petrok, C.N. & Esteban, J.A.

(2010) PTEN is recruited to the postsynaptic terminal for NMDA receptor- dependent long-term depression.EMBO J.,29, 2827–2840.

Krueger, D.D. & Bear, M.F. (2011) Toward fulﬁlling the promise of molecu- lar medicine in fragile X syndrome.Annu. Rev. Med.,62, 411–429.

(9)

Lauterborn, J.C., Rex, C.S., Kramar, E., Chen, L.Y., Pandyarajan, V., Lynch, G. & Gall, C.M. (2007) Brain-derived neurotrophic factor rescues synaptic plasticity in a mouse model of fragile X syndrome. J. Neurosci., 27, 10685–10694.

Mao, L., Jia, J., Zhou, X., Xiao, Y., Wang, Y., Mao, X., Zhen, X., Guan, Y., Alkayed, N.J. & Cheng, J. (2013) Delayed administration of a PTEN inhibitor BPV improves functional recovery after experimental stroke.

Neuroscience,231, 272–281.

Meredith, R.M., Holmgren, C.D., Weidum, M., Burnashev, N. & Mansvel- der, H.D. (2007) Increased threshold for spike-timing-dependent plasticity is caused by unreliable calcium signaling in mice lacking fragile X gene FMR1.Neuron,54, 627–638.

Michalon, A., Sidorov, M., Ballard, T.M., Ozmen, L., Spooren, W., Wett- stein, J.G., Jaeschke, G., Bear, M.F. & Lindemann, L. (2012) Chronic pharmacological mGlu5 inhibition corrects fragile X in adult mice. Neu- ron,74, 49–56.

Mientjes, E.J., Nieuwenhuizen, I., Kirkpatrick, L., Zu, T., Hoogeveen- Westerveld, M., Severijnen, L., Rife, M., Willemsen, R., Nelson, D.L. &

Oostra, B.A. (2006) The generation of a conditional Fmr1 knock out mouse model to study Fmrp function in vivo.Neurobiol. Dis., 21, 549–

555.

Osterweil, E.K., Kind, P.C. & Bear, M.F. (2012) Lifting the mood on treat- ing fragile X.Biol. Psychiat.,72, 895–897.

Osterweil, E.K., Chuang, S.C., Chubykin, A.A., Sidorov, M., Bianchi, R., Wong, R.K. & Bear, M.F. (2013) Lovastatin corrects excess protein synthesis and prevents epileptogenesis in a mouse model of fragile X syndrome.Neuron,77, 243–250.

Pan, F., Aldridge, G.M., Greenough, W.T. & Gan, W.B. (2010) Dendritic spine instability and insensitivity to modulation by sensory experience in a

mouse model of fragile X syndrome. Proc. Natl. Acad. Sci. USA, 107, 17768–17773.

Pfeiffer, B.E. & Huber, K.M. (2007) Fragile X mental retardation protein induces synapse loss through acute postsynaptic translational regulation.

J. Neurosci.,27, 3120–3130.

Pfeiffer, B.E., Zang, T., Wilkerson, J.R., Taniguchi, M., Maksimova, M.A., Smith, L.N., Cowan, C.W. & Huber, K.M. (2010) Fragile X mental retardation protein is required for synapse elimination by the activity-dependent transcription factor MEF2.Neuron,66, 191–197.

Schmid, A.C., Byrne, R.D., Vilar, R. & Woscholski, R. (2004) Bisperox- ovanadium compounds are potent PTEN inhibitors.FEBS Lett.,566, 35– 38.

Scotto-Lomassese, S., Nissant, A., Mota, T., Neant-Fery, M., Oostra, B.A., Greer, C.A., Lledo, P.M., Trembleau, A. & Caille, I. (2011) Fragile X mental retardation protein regulates new neuron differentiation in the adult olfactory bulb.J. Neurosci.,31, 2205–2215.

Seese, R.R., Babayan, A.H., Katz, A.M., Cox, C.D., Lauterborn, J.C., Lynch, G. & Gall, C.M. (2012) LTP induction translocates cortactin at distant synapses in wild-type but not Fmr1 knock-out mice. J. Neurosci., 32, 7403–7413.

Sharma, A., Hoeffer, C.A., Takayasu, Y., Miyawaki, T., McBride, S.M., Klann, E. & Zukin, R.S. (2010) Dysregulation of mTOR signaling in fragile X syndrome.J. Neurosci.,30, 694–702.

Stoppini, L., Buchs, P.A. & Muller, D. (1991) A simple method for organotypic cultures of nervous tissue.J. Neurosci. Meth.,37, 173–182.

Tang, S.J., Reis, G., Kang, H., Gingras, A.C., Sonenberg, N. & Schuman, E.M. (2002) A rapamycin-sensitive signaling pathway contributes to long- term synaptic plasticity in the hippocampus.Proc. Natl. Acad. Sci. USA, 99, 467–472.