Seminars in Cell & Developmental Biology 22 (2011) 408– 415

Contents lists available at ScienceDirect

Seminars in Cell & Developmental Biology

j ourna l ho me pag e: www.elsev ier .com/ locate /semcdb

Review

Synapsins: From synapse to network hyperexcitability and epilepsy

Anna Fassioa, Andrea Raimondib, Gabriele Lignania,b, Fabio Benfenati a,b, Pietro Baldelli a,b,∗

a Department of Experimental Medicine, Section of Physiology and National Institute of Neuroscience, University of Genova, Genova, Italyb Department of Neuroscience and Brain Technologies, The Italian Institute of Technology Central Laboratories, Genova, Italy

a r t i c l e i n f o

Article history:

Available online 26 July 2011

Keywords:

Synaptic transmission

Knockout mice

Synaptic vesicles

Seizure

Epileptogenesis

a b s t r a c t

The synapsin family in mammals consists of at least 10 isoforms encoded by three distinct genes and

composed by a mosaic of conserved and variable domains. Synapsins, although not essential for the basic

development and functioning of neuronal networks, are extremely important for the fine-tuning of SV

cycling and neuronal plasticity.

Single, double and triple synapsin knockout mice, with the notable exception of the synapsin III

knockout mice, show a severe epileptic phenotype without gross alterations in brain morphology and

connectivity. However, the molecular and physiological mechanisms underlying the pathogenesis of the

epileptic phenotype observed in synapsin deficient mice are still far from being elucidated. In this review,

we summarize the current knowledge about the role of synapsins in the regulation of network excitabil-

ity and about the molecular mechanism leading to epileptic phenotype in mouse lines lacking one or

more synapsin isoforms. The current evidences indicate that synapsins exert distinct roles in excitatory

versus inhibitory synapses by differentially affecting crucial steps of presynaptic physiology and by this

mean participate in the determination of network hyperexcitability.

© 2011 Elsevier Ltd. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4082. Differential expression and localization of synapsins in excitatory and inhibitory synapses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4093. Role of synapsins in SVs cycling assessed by live cell imaging techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4094. Different presynaptic role of synapsins at excitatory and inhibitory synapses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4105. From the alteration of single synaptic contact to the network hyperexcitability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4126. Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 413

Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 413References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 414

1. Introduction

The analysis of synapsin knock-out (KO) mouse lines has clearly

shown that the synapsins (Syns) are involved in the regulation of

the excitability of neuronal networks, and that impairment of Syn

function can result in the onset of pathological conditions. In fact,

SynI−/−, SynII−/− SynI,II−/− and SynI,II,III−/− but not SynIII−/−mice are all prone to epileptic seizures, which start to develop

approximately at two months of age, and progressively aggravate

with aging and with the number of Syn genes ablated [1,2].

∗ Corresponding author at: Department of Experimental Medicine, University of

Genova, Viale Benedetto XV, 3, 16132 Genova, Italy. Tel.: +39 010 353 8186;

fax: +39 010 353 8194.

E-mail address: pietro.baldelli@unige.it (P. Baldelli).

The phenotype of the various Syn KO mice became even more

interesting after the discovery of epileptogenic mutations of SYN

genes in human. Genetic analyses in human populations have iden-

tified a nonsense mutation in the gene coding for SynI, likely to

cause mRNA decay, as the cause of epilepsy in a family with history

of epilepsy alone, or associated with aggressive behavior, learn-

ing disabilities or autism [3]. Very recently, a nonsense mutation

in SYN1 gene was identified in all affected individuals from a large

French–Canadian family segregating epilepsy and autism spectrum

disorders (ASDs) and additional missense mutations were found

in 1.0% and 3.5% of French–Canadian individuals with ASDs and

epilepsy, respectively [4]. In addition, genetic mapping analysis

identified variations in the SYN2 gene as significantly contributing

to epilepsy predisposition [5,6].

A recent study has analyzed and classified the types of seizures

affecting SynI,II−/− mice, as well as the transitions between the

1084-9521/$ – see front matter © 2011 Elsevier Ltd. All rights reserved.doi:10.1016/j.semcdb.2011.07.005

A. Fassio et al. / Seminars in Cell & Developmental Biology 22 (2011) 408– 415 409

different seizure elements. In this work, the authors describe three

different clusters of epileptic activity: the truncus dominated clus-

ter, the myoclonic cluster, and the running fit cluster. The first type

of epileptic activity is characterized by a highly conserved sequence

of elements, which likely reflects the sequential activation of spe-

cific neuronal populations. The two other clusters are instead much

more variable, probably indicating a more random activation of

neurons in multiple brain areas [7]. This study indicates how lack

of Syn proteins can generate epileptic seizures by evoking a range

of complex mechanisms, involving distinct neuronal populations

in various brain areas.

Here we analyze and summarize various experimental evi-

dences describing the neurophysiological mechanisms underlying

the onset of seizures in Syn KO mouse models that will hopefully

help the understanding of the molecular processes leading to the

development of epilepsy.

2. Differential expression and localization of synapsins inexcitatory and inhibitory synapses

In the central nervous system, the vast majority of nerve termi-

nals express at least one Syn isoform and Syns are concentrated in

presynaptic boutons and associated exclusively with small synap-

tic vesicles (SVs) [8,9]. Syn antibodies have been extensively used

as reliable synaptic markers; however, already from the pioneer

study on SynI and II distribution in the mammalian central nervous

system, differential expression of Syn isoforms in subset of cen-

tral synapses was observed. Greengard and coworkers, analysing

rat brain frozen section by standard immunofluorescence, revealed

that while hippocampal CA3 mossy fiber terminals expressed both

SynI and II isoforms, in the deep cerebellar nuclei SynIIa was not

detectable from Purkinje cell GABAergic terminals, and SynIa and

IIb were expressed at lower level with respect to SynIb. In the

nucleus of the trapezoid body of the brainstem, SynIb was equally

present in all nerve terminals but SynIIb was present only in few

synapses [9]. In the rat retina SynI and II are absent in all ribbon

synapses and differentially distributed in the conventional synaptic

terminals of amacrine cells [10,11]. In the olfactory bulb, whereas

the core region expresses comparable level of SynI and II, the sur-

face region had significantly higher levels of SynIIa comparing to

SynI [12]. Axon terminals in the dorsal lateral geniculate nucleus of

murine brain also revealed differences in term of SynI/II expression:

glutamatergic cortical inputs express both isoforms, GABAergic

synapses express only SynI while both isoforms were absent in

the excitatory terminals arising from the retina [13]. Distribution

of Syns I and II in glutamatergic and GABAergic murine terminals

was also analyzed, revealing a complete colocalization of the two

isoforms in VGlut1-positive terminals of the stratum lucidum as

well as in VGlut2-positive terminals in the striatum. Also GABAergic

VGAT-positive terminals in the striatum appeared to express both

SynI and II, whereas VGAT-containing terminals in CA3 pyramidal

cell layer and stratum lucidum of the hippocampus did contain nei-

ther SynI nor SynII [14]. In a subsequent study performed in the rat

neocortex, 30% and 50% of VGAT positive terminals were also found

to partially express SynI and SynII, respectively [15]. In the same

paper it was shown that virtually all excitatory synapses positive for

VGlut1 express both SynI and SynII, whereas 30% of VGlut2-positive

puncta express SynI, another 30% express SynII and a significant

portion of VGlut2 positive terminals are negative for either Syn

isoform [15]. However, a recent study, in which the composition of

glutamatergic and GABAergic synapses of mouse somato-sensory

cortex has been analyzed by array tomography, showed that vir-

tually all synapses are recognized by anti-SynI antibody, while

antibodies to other synaptic proteins revealed the existence of sev-

eral synaptic subtypes [16]. The SynI content, however, varied in

intensity depending on the synapse type. In agreement with the

work of Bragina, VGlut1-positive excitatory synapses and VGAT-

positive inhibitory synapses expressed the highest and the lowest

level of SynI, respectively. The presence of lower levels of expres-

sion in GABAergic terminals was previously revealed also for other

synaptic proteins as SV2s, synaptotagmins, syntaxins and synapto-

physins [15,17,18] and it is possibly in correlation with the different

functional properties of the synapses.

The distribution of SynIII in the adult mouse forebrain was also

examined and found to be significantly different from the distribu-

tion pattern seen for SynI and II [19]. The levels of SynIII in nerve

terminals were much lower than those of Syns I and II. Moreover,

differentially from SynI and II, SynIII was also highly expressed in

the cell body and processes of immature neurons in neurogenic

regions of the adult brain, such as the hippocampal dentate gyrus,

the rostral migratory stream, and the olfactory bulb and was found

to play a role in neural progenitor cell development in the adult

hippocampus [20].

Considering that most of the functional studies on SV cycling

and neurotransmission performed on Syn KO mice (see below) used

hippocampal primary culture as an experimental model, we here

analyzed the differential expression of Syn isoforms in excitatory

and inhibitory synapses from cultured wild-type hippocampal neu-

rons. As shown in Fig. 1, we revealed a significant lower level of

colocalization for all Syn isoforms with inhibitory (VGAT-positive)

synapses as compared to excitatory (VGlut1-positive) synapses and

a lower level of expression of all three synapsins in GABAergic

versus glutamatergic synapses. These data suggest that inhibitory

synapses from hippocampal cultures alternatively express one Syn

isoform or that a significant portion of GABAergic synapses is

devoid of Syn, although we cannot exclude that the level of expres-

sion of the various Syn isoforms is below the detection threshold

of our analysis.

The impact of the various Syn isoforms on synaptic physiology

is also dependent on their developmental expression. The onset

of Syns I and II expression coincides with the time of commitment

from progenitor cells to differentiated neurons, and it is particularly

high during synaptogenesis [8,21–25]. In cultured hippocampal

neurons, the expression of Syns I and II progressively raises with

time, whereas SynIII is highly expressed in the first week in culture

during active process elongation, and is enriched in cell bodies, as

well as in growth cones [26–28].

Taken together, immunolocalization data reveal distinct local-

ization patterns among Syn isoforms. Particularly interesting for

the goal of this review is the differential localization among excita-

tory and inhibitory synapses which suggest a different role of Syns

in GABAergic versus glutamatergic synapses. The overall outcome

of the papers investigating on Syn expression in central synapses,

using different techniques and experimental models, suggests that

synapsins are indeed less expressed in inhibitory versus excitatory

terminals. The difference in expression has been revealed at the

level of the cerebral cortex and the hippocampus and could pos-

sibly explain the hyperexcitability observed in neuronal networks

lacking one or more SYN genes that lead to epileptic seizures in Syn

KO adult animals.

3. Role of synapsins in SVs cycling assessed by live cellimaging techniques

Since the development of the optical technique to assay presy-

naptic function by quantitative fluorescence imaging of FM-dyes

[29], studies on the role of Syns in the regulation of SV cycling

by in vivo imaging have been performed. FM dyes, as well as the

more recent genetically engineered probes of the synaptopHluo-

rin family [30,31], are optimal tools for the study of Syn function

as they allow an exquisite presynaptic analysis at the level of

410 A. Fassio et al. / Seminars in Cell & Developmental Biology 22 (2011) 408– 415

Fig. 1. Distribution of synapsin I (A), synapsin II (B) and synapsin III (C) in excitatory and inhibitory synapses. (a) Cultured hippocampal neurons (14 DIV) co-immunostained

for synapsin isoforms (green) with excitatory and inhibitory markers (VGlut1 (red) and VGAT (blue), respectively). Bars, 10 �m. (b) The extent of colocalization between

VGlut1/synapsins and VGAT/synapsins is shown as percentage of colocalizing pixels (Mander’s coefficient) (***p < 0.001 Student’s t-test). (c) Relative frequency distribution

of synapsin fluorescence intensity in VGlut1- or VGAT-positive puncta (***p < 0.001 Kolmogorov–Smirnov test).

small central synapses. Using FM1-43, Ryan et al. showed that

both the number of vesicles exocytosed during brief action poten-

tial trains and the total recycling vesicle pool are significantly

reduced in SynI−/− mice, while the kinetics of endocytosis and

SV repriming appeared normal [32]. An opposite phenotype using

the same FM dye was described in SynIII deficient boutons, where

the total recycling pool of SVs was reported to be higher com-

pared to wild type boutons [33]. Using FM4-64, a red shifted

variant of FM1-43, the kinetics of SV turnover were correlated

with the rate of dispersion of SynIa. Dispersion of SynI appeared

to track the SV turnover rate linearity and to be tightly regu-

lated by phosphorylation. Calcium/calmodulin dependent kinase

(CaMK) phosphorylation controlled SV mobilization at low stim-

ulus frequency, while mitogen-activated protein kinase (MAPK)

phosphorylation appeared critical at both low and high stimu-

lus frequencies [34,35]. Calcium influx during depolarization also

activated PKA phosphorylation at Syn site-1 and SynI PKA phos-

phorylation has been reported to play a major role in tuning the

size of the recycling pool in response to depolarization [36]. Opti-

cal recordings from SynI,II−/− hippocampal neurons revealed a role

of both SynI and/or SynII in the mechanism that ensures function-

ally effective allocation of a limited number of SVs during synapse

formation, with a significant reduction of the recycling pool in

synapses lacking both SynI and SynII [37]. FM dye experiments

have also been performed on SynI,II,III−/− hippocampal neurons

to study the role of Syns in the maintenance and regulation of the

ready releasable pool (RRP) and reserve pool (RP) of SVs [38]. This

study revealed that the loss of all Syns produced a reduction in the

total recycling pool attributable to a significant impairment at the

level of the RP. The latter effect was partially rescued by expres-

sion of SynIIa, but the rescuing ability of the other Syn isoforms has

not been tested [38]. SynIa dynamics has been reported to be mod-

ulated by the presynaptic scaffold protein piccolo. The rate of SV

cycling was enhanced in synapses lacking piccolo due to an alter-

ation in the activity-dependent dynamics of SynIa at the terminals

[39]. A defect in SV recycling has also been revealed by FM4-64 live

imaging experiments from ATM deficient cortical neuron cultures.

ATM is an ubiquitous nuclear PI3-kinase and ATM deficiency cause

ataxia-telangiectasia. ATM has been reported to be expressed also

at the level of the cytoplasm in brain cells and to selectively bind

phosphorylated Syn and synaptobrevin [40].

SynaptophysinpHluorin (SypHy) [41] transfected hippocampal

primary cultures from SynI−/− mice embryos have been used to

evaluate the effect of SynIa tyrosine phosphorylation on SV dynam-

ics. The tyrosine-dephosphomimetic mutant of SynI increases the

RRP and the RP of SVs suggesting that tyrosine phosphorylation,

opposite to serine phosphorylation, favors recovery and reconsti-

tution of SV stores at the terminals [42]. Recently, another kinase

phosphorylating SynI, cdk5, was reported to negatively modulate

the recycling pool of SVs, although whether this effect is mediated

by phosphorylation of SynI or of other nerver terminal substrates

is not clear [43]. Finally, SypHy assay has been used to evaluate

the effect of human SynI mutation associated with ASDs and/or

epilepsy revealing impairment in the size and trafficking of synaptic

vesicle pools when compared to synapses expressing hSynI [4].

All the reported data support a direct involvement of Syns in

the regulation of SV dynamics at the synapses, with an emergent

view in which the absence of Syns decreases the total number of

recycling SV at the nerve terminals. However, most of the data

collected so far by directly assessing the presynaptic role of Syns

using fluorescence techniques (either FM-dyes or synaptopHluo-

rins) have been performed in hippocampal primary cultures with

no discrimination between excitatory and inhibitory terminals. Due

to the differential expression of Syns (see above) and different size

of SV pools in glutamatergic versus GABAergic terminals [44], fur-

ther experiments are required which combine the optical technique

based on SV fluorescent probes with immunostaining, to distin-

guish inhibitory from excitatory terminals.

4. Different presynaptic role of synapsins at excitatory andinhibitory synapses

Central synapses in mice lacking one or more Syn isoforms,

with the notable exception of SynIII−/− KO mice, display a marked

A. Fassio et al. / Seminars in Cell & Developmental Biology 22 (2011) 408– 415 411

decrease in SV density, reflecting a depletion of the RP [1,2,45–47].

This is consistent with a well described pre-docking role of the Syns

in the assembly of the RP of SVs through SV clustering and actin

binding [48–50]. In addition, hippocampal SynI−/− neurons suffer

of a reduced size of the recycling pool and a decreased recruit-

ment of recycling SVs to the RRP during synapse maturation [32,37].

It is possible that the enhancement of synaptic depression during

repetitive stimulation contributes to seizure development, possibly

because of a different pre-docking role that Syns exert at excitatory

and inhibitory synapses. Most of the data show that Syn dele-

tion enhances synaptic depression during repetitive stimulation at

larger extent at excitatory synapses than inhibitory synapses [46].

Walaas and coworkers showed that deletion of SynsI and II induced

higher synaptic depression, reduced post-tetanic potentation (PTP)

and decreased facilitation in excitatory synapses from thalamo-

cortical neurons of the dorsal lateral geniculate nucleus [13] and

in CA3–CA1 hippocampal synapses [51–53]. The comparison of

excitatory and inhibitory autapses, performed in wild type (WT)

and SynI,II,III−/− mice, further confirmed that the enhancement of

synaptic depression mainly involves the excitatory transmission,

leaving GABAergic synapses almost unaffected [46].

However in SynI−/− mice, this scenario appeared to change

significantly. Cultured hippocampal neurons from SynI−/− mice

showed inhibitory synapses becoming easily fatigued upon

repeated application of hypertonic sucrose and recovering slowly

from depression. Upon stimulation, presynaptic terminals showed

a decrease in the number of SVs in the RP but not in the RRP, which

was slightly more pronounced in GABAergic terminals than in glu-

tamatergic ones [54]. More recently, a moderate increase of synap-

tic depression evoked by trains of action potentials was observed

at inhibitory synapses of cultured SynI−/− hippocampal neurons.

This phenotype, scarcely detectable at low stimulation frequencies,

became clear-cut at higher frequencies, and was accompanied by

a slow-down of recovery from depression [55]. However, different

results can be obtained changing brain area. For example, cultured

cortical neurons from SynI−/− mice showed no signs of increased

synaptic depression upon sustained repetitive stimulation at both

excitatory and inhibitory autapses [56]. In conclusion, an alteration

of the predocking actions of Syns is not sufficient to define the

cause of the epileptic phenotype, although it is possible to speculate

that an imbalance between excitatory and inhibitory transmission

could arise from the high frequency firing characterizing GABAer-

gic interneurons, that makes GABA release particularly sensitive to

the relative SV depletion induced by Syn deletion.

To obtain a more complete picture it is necessary to take into

consideration the post-docking activity of the Syns that appear to

be different affect in glutamatergic versus GABAergic synapses. A

growing amount of convergent results obtained from three inde-

pendent research groups revealed a reduction of eIPSC amplitude in

response to a single isolated stimulus in hippocampal and cortical

SynI−/−, SynIII−/− and SynI,II,III−/− neurons [33,46,54–56].

In hippocampal autapses from SynI,II,III−/− animals, eIPSC

amplitude undergoes a 30% reduction, with no effects on frequency

and amplitude of mIPSCs. In agreement with this functional char-

acterization, electron microscopy analysis revealed that the SV

loss was not restricted to the RP, as normally observed in excita-

tory synapses derived from SynI −/−, SynIII−/− and SynI,II,III−/−mice [1,33,46], but it involved mainly the docked SVs in inhibitory

synapses [46]. A similar effect was observed in inhibitory terminals

of SynI−/− cultured hippocampal neurons [55]. Here, GABAergic

transmission showed a normal spontaneous release, paired-pulse

depression and PTP, but SynI deletion clearly decreased the eIPSCs

amplitude, an effect that became apparent only in fully differenti-

ated neurons. This effect was shown to be due to a decrease in the

size of RRP, rather than by changes in release probability or quan-

tal size. The reduction of the RRP was caused by a decrease in the

number of SVs released by single synaptic bouton in response to

the action potential, in the absence of variations in the number of

synaptic contacts between couples of monosynaptically connected

neurons. Subsequently, eIPSC amplitude was studied in SynI−/−cortical autapses and such reduction of inhibitory transmission was

confirmed [56]. The decrease in the mean amplitude of eIPSCs was

not accompanied by any change in the amplitude of the mIPSCs,

indicating the absence of postsynaptic effects and/or alterations in

the quantum content. As observed in hippocampal neurons, also

in cortical neurons the cumulative amplitude analysis suggested

that this effect is exclusively the consequence of a decreased size

of the RRP, consistent with the fact that paired-pulse depression of

GABAergic transmission was not affected by the absence of SynI,

indicating the absence of changes in release probability.

Surprisingly, in the same experimental substrate, an enhance-

ment of the eEPSC amplitude was revealed [56]. Interestingly, such

effect was here observed also in hippocampal excitatory autapses

obtained from SynI−/−mice (Fig. 2) and in Shaffer collateral-to-CA1

Fig. 2. Differential effects of synapsins in hippocampal excitatory and inhibitory

neurons. (A) Representative phase-contrast of WT (left) and SynI−/− (right) autaptic

neurons and the corresponding traces of autaptic eEPSCs (B) and eIPSCs (C). The

analysis of the mean (±SEM) amplitude of eEPSCs and eIPSCs (right panels in B

and C, respectively) reveals a primary excitatory/inhibitory imbalance associated

with SynI deletion (S. Casagrande, F. Benfenati and P. Baldelli). WT (black bars) and

SynI−/− (red bars). Scale bars: A, 40 �m; B and C, amplitude = 200 pA (B) and 400 pA

(C), time = 30 ms (B) and 40 ms (C).

412 A. Fassio et al. / Seminars in Cell & Developmental Biology 22 (2011) 408– 415

excitatory synapses obtained from SynI,II,III−/− mice (unpublished

data from our laboratory). In both cortical [56] and hippocampal

(Fig. 2) neurons we found that the increase in the amplitude of

eEPSC was not associated with an increase in the mEPSC ampli-

tude. As observed for GABAergic transmission, the lack of changes

in miniature events indicates that, in the absence of postsynaptic

effects and/or changes in the mean quantum content of SVs, the

increase in eEPSC amplitude depends on an increased number of

SVs released in response to the action potential. This effect seems to

be entirely attributable to an increased size of the RRP, rather than

to an increase in release probability, as suggested by the cumulative

amplitude analysis.

These results clearly show that SynI, exerts completely oppo-

site post-docking actions on inhibitory and excitatory synapses that

are respectively impaired and enhanced by the deletion of the pro-

tein. The logic consequence of these changes is the generation of

an unbalance between excitatory and inhibitory inputs, that could

constitute the basis for the development of an initial state of hyper-

excitability, progressively evolving in the onset of the first seizures.

While the generation of the epileptic phenotype can be understood

on the basis of these synaptic phenotypes, it is more complex to

understand the molecular mechanisms by which deletion of the

very same presynaptic protein triggers so different, often opposite,

effects on the function of excitatory and inhibitory terminals.

A possible explanation derives from two observations. First, it is

known that glutamatergic neurons exhibit a large percentage of SVs

which are resistant to release due to a lower probability of release

compared to GABAergic neurons, suggesting distinct mechanisms

of SV trafficking and recruitment for release [1,44]. Second, we and

other groups [15,16] have shown that the level of expression of SynI

and II in excitatory synapses is higher than in inhibitory terminals.

Although an increase in the RRP size in mutant excitatory synapses

is not consistent with the very limited and non-significant changes

in the number of docked SVs estimated by electron microscopy

(EM) in glutamatergic terminals lacking Syns [1,2,46,47], cumula-

tive amplitude analysis gives a functional and dynamic estimation

of the RRP size in contrast with the morphological description of

RRP provided by electron microscopy. Therefore, it is possible that

the observed increase in the functional RRP size in mutant gluta-

matergic terminals is attributable to the partial disassembly of the

RP which may recruit more SVs to the RRP.

From this complex scenario, it appears that in neurons where

one or more Syns are deleted, glutamatergic transmission is

impaired during sustained high frequency activity, but it is fully

preserved (SynI,II,III−/−) or also enhanced (SynI−/−) in response

to single action potentials. On the other hand, GABAergic transmis-

sion is clearly defective during basal electrical activity and slightly

reduced during high-frequency stimulation. The analysis of the

pre-docking action of Syns on excitatory and inhibitory neurons

indicates that, although Syn ablation has probably a stronger effect

on the depression of excitatory synapses, the milder effect observed

in inhibitory synapses could have a higher impact on network

excitability. This might be due to the bursting-type firing proper-

ties that characterize inhibitory neurons, which make them more

sensitive to SVs depletion.

On the other hand, when focusing on the post-docking role of

Syns, it is clear that Syn deletion mainly impairs the inhibitory

input, reducing the strength of basal GABAergic transmission, leav-

ing basal excitatory transmission unaffected or even enhanced

[46,54–56]. Under this condition, not only the phasic, GABAA

dependent synaptic inhibition is impaired, but potentially also

the tonic extrasynaptic current due to GABA spillover and/or

GABAB-mediated inhibition. This multiple impairment of complex

inhibitory mechanisms may force neuronal circuits into a state

of basal heightened excitability facilitating the spontaneous or

stimulus-evoked onset of epileptic seizures.

5. From the alteration of single synaptic contact to thenetwork hyperexcitability

Syns are therefore involved in crucial steps of presynaptic phys-

iology at both excitatory and inhibitory synapses. Thus, it became

fundamental to understand how and at which extent Syns can par-

ticipate in the determination of the firing activity of the neuronal

network and in its change towards a state of hyperexcitability.

To this aim, we have taken advantage from the use of multisite

extracellular recordings performed using multielectrodes arrays

(MEAs) made up of 60 planar microelectrodes (Fig. 3A and B). MEA

recordings have been used for investigating whether SynI−/− cor-

tical networks exhibit developmentally regulated hyperexcitability

in vitro [56]. Primary cortical cultures from WT mice exhibited

a spontaneous electrical activity characterized by both isolated

random spikes and bursting activity. The activity of the WT net-

work increased with the functional maturation in vitro, and finally

reached a steady-state at the third week in vitro (Fig. 3C). In con-

trast, SynI−/− cultures displayed a constant increase in firing rate,

that never reached a steady-state level, becoming progressively

more marked with network maturation. Interestingly, the progres-

sive build-up of hyperexcitability observable in SynI−/− networks

during in vitro development parallels with synaptic maturation

and coincides with the period in which WT neurons increase SynI

expression [26]. This late aggravation of the hyperexcitability is also

consistent with the late appearance of the impairment in eIPSCs

observed in cultured hippocampal neurons obtained from SynI−/−mice [55], showing a intriguing temporal coherence with the onset

of the first epileptic seizures in SynI−/− mice, starting at 2 months

of postnatal development and becoming progressively more severe

with age [2,57].

The block of GABAergic transmission, by bicuculline, only

slightly attenuated the difference in firing rate between WT and

SynI−/− networks (Fig. 3D). This result shows that the hyperex-

citability of SynI−/− networks, is not exclusively due to a decrease

in GABAergic transmission [54,55], but also to a concomitant

increase in glutamatergic transmission. The absence of SynI leads

to a primary imbalance between inhibitory and excitatory systems,

thereby generating diffuse spontaneous and stimulation-evoked

hyperexcitability. Such an imbalance could be subsequently main-

tained, and possibly aggravated, by the development of long-term

maladaptive plasticity processes induced by the intense and repet-

itive firing.

Extracellular MEA recordings were also used to characterize

the excitability and the generation of epileptic paroxysms at cel-

lular and network levels in acute hippocampal/limbic cortex slices

from presymptomatic and symptomatic TKO mice and to uncover

changes in epileptic propensity over age [58]. Because of the spo-

radic presence of spontaneous epileptiform activity in SynI,II,III−/−slices, the investigation was carried out by treating slices with

4-aminopyridine (4AP) that enhances both glutamatergic and

GABAergic transmissions. Treatment with 4AP induced interictal

(I-IC) events in the hippocampus and limbic cortex, sporadically

interrupted by long-lasting cortical ictal (IC) discharges in both WT

and SynI,II,III−/− slices. We found that the incidence of I-IC and

IC like activity was already much more intense in presymptomatic

SynI,II,III−/− than in WT slices. In particular, while IC discharges

were confined to cortical regions in WT slices, hippocampal IC

seizures, were also present in SynI,II,III−/− slices. Interestingly, the

cortical area involved by the spread of IC discharges was increased

in SynI,II,III−/− slices, suggesting an impairment of the inhibitory

mechanisms which should circumscribe the areas of localized

hyperexcitability. The higher frequency of these paroxysms and

their larger spread in slices already observed in presymptomatic

SynI,II,III−/− mice demonstrate that an alteration of network

excitability is already present and that a progressive aggravation

A. Fassio et al. / Seminars in Cell & Developmental Biology 22 (2011) 408– 415 413

Fig. 3. Spontaneous activity of cortical cultures from WT and SynI−/− mice. (A and B) A WT network cultured over a MEA for 31 DIV (bar: 250 and 30 �m in panels A

and B, respectively). (C) Primary cortical neurons from WT (gray bars) and SynI−/− (black bars) mice were cultured onto MEAs for various DIV and monitored at various

developmental stages. Firing rates calculated for each experimental group, are shown as means (±SEM) at the following developmental stages: 12–15, 18–20, 24–26, and

31–35 DIV. **p < 0.01 between genotype within each developmental stage; Mann–Whitney U-test for independent samples. (D) The effects of BIC on the firing rate of

the MEA cultures (24–35 DIV) are shown in the bar diagram as means (±SEM) for WT (black bars) and SynI−/− (red bars) neurons; **p < 0.01 SynI−/− versus WT within

treatment;◦◦p < 0.01 BIC versus Basal within genotype; Kruskal–Wallis multiple comparison test.

Modified from Chiappalone et al. [56].

of the condition of hyperexcitability anticipates the onset of the

first epileptic symptoms at an age of about 2/3 months. It is now

fundamental to study and define the epileptogenic processes that

drive the neuronal network to the late onset of the first spontaneous

seizure and rapidly evolves in an overt full-blown epileptic.

6. Concluding remarks

Genetic ablation of SYNI, SYNII, SYNI/II and SYNI/II/III genes in

mice all resulted in an epileptic phenotype. Moreover SYNs are the

only SV protein genes whose mutations were found to be associ-

ated with human epilepsy [3–6]. The genetic deletion of the Syns in

mice is of particular interest in this respect, because it represents

the first mouse model of human epilepsy based on an alteration

of SV proteins. Syn appear to be highly expressed in the central

nervous system and to regulate both neuronal development and

neurotransmission. Here, we have summarized the current knowl-

edge of the molecular mechanisms leading to hyperexcitability in

neuronal networks missing one or more Syn isoforms.

From the current evidence it appears that Syns are differentially

expressed and exert distinct functions at the level of inhibitory

and excitatory terminals. In particular, Syns is more abundant

in hippocampal and cortical excitatory synapses, while they are

expressed at lower levels in inhibitory terminals. The loss of Syns

was shown to disrupt the RP of SVs and enhance synaptic depres-

sion more severely in excitatory than in inhibitory synapses. On the

contrary, neurotransmitter release in response to a single presy-

naptic stimulus is strongly reduced in GABAergic terminals and

not affected, or even enhanced, in glutamatergic boutons. In this

context it is possible that the imbalance between excitatory and

inhibitory inputs could mainly originate from the impairment of

GABAergic transmission during both basal electrical activity and

high-frequency stimulation. Under this condition, both the pha-

sic synaptic and the tonic extrasynaptic inhibition due to GABA

spillover will be decreased possibly forcing neuronal circuits into

a state of basal heightened excitability which facilitates epileptic

seizures.

However, despite the vast number of studies, many aspects of

how Syns are able to affect the neuronal network excitability still

remain to be elucidated. Considering that the first seizures start

to be detectable at the SynI/II expression peak, which coincides

with synapse maturation and refinement, it will be fundamen-

tal to follow how the changes of crucial steps of both excitatory

and inhibitory neurosecretion processes, due to Syn deletion, are

developmentally modulated. The possible deficits in the period

of synapse rearrangement and refinement may also explain the

recently reported association of SYN1 mutations with ASDs [4]. In

this context, Syn KO mice provide an alternative and innovative

model of human epilepsy useful to study the cellular mechanisms

underlying genesis of seizures and to follow in detail the progres-

sion of the epileptogenesis process.

Acknowledgements

We thank Drs. Paul Greengard (The Rockefeller University, New

York, NY) and Hung-Teh Kao (Brown University, Providence, RI) for

414 A. Fassio et al. / Seminars in Cell & Developmental Biology 22 (2011) 408– 415

the long-standing collaboration and for invaluable and stimulat-

ing discussions. Work in the Authors’ laboratories was supported

by research grants from Telethon, Italy (Grants GGP05134 and

GGP09134 to FB and Grant GGP09066 to PB), the Compagnia di

San Paolo Torino (PB, AF and FB) and the Italian Ministry of Health

(to PB and AF).

References

[1] Li L, Chin L, Shupliakov O, Brodin L, Sihra TS, Hvalby Ø, et al. Impairmentof synaptic vesicle clustering and of synaptic transmission, and increasedseizure propensity, in synapsin I-deficient mice. Proc Natl Acad Sci USA1995;92:9235–9.

[2] Rosahl TW, Spillane D, Missler M, Herz J, Selig DK, Wolff JR, et al. Essen-tial functions of synapsins I and II in synaptic vesicle regulation. Nature1995;375:488–93.

[3] Garcia CC, Blair HJ, Seager M, Coulthard A, Tennant S, Buddles M, et al. Identi-fication of a mutation in synapsin I, a synaptic vesicle protein, in a family withepilepsy. J Med Genet 2004;41:183–6.

[4] Fassio A, Patry L, Congia S, Onofri F, Piton A, Gauthier J, et al. SYN1 loss-of-function mutations in ASD and partial epilepsy cause impaired synapticfunction. Hum Mol Genet 2011;(March) [Epub ahead of print].

[5] Cavalleri GL, Weale ME, Shianna KV, Singh R, Lynch JM, Grinton B, et al. Multi-centre search for genetic susceptibility loci in sporadic epilepsy syndrome andseizure types: a case–control study. Lancet Neurol 2007;6:970–80.

[6] Lakhan R, Kalita J, Misra UK, Kumari R, Mittal B. Association of intronic polymor-phism rs3773364 A>G in synapsin-2 gene with idiopathic epilepsy. Synapse2010;64:403–8.

[7] Etholm L, Heggelund P. Seizure elements and seizure element transitions dur-ing tonic–clonic seizure activity in the synapsin I/II double knockout mouse: aneuroethological description. Epilepsy Behav 2009;14:582–90.

[8] De Camilli P, Harris Jr SM, Huttner WB, Greengard P. Synapsin I (protein I), anerve terminal-specific phosphoprotein. II. Its specific association with synap-tic vesicles demonstrated by immunocytochemistry in agarose-embeddedsynaptosomes. J Cell Biol 1983;96:1355–73.

[9] Sudhof TC, Czernik AJ, Kao H, Takei K, Johnson PA, Horichi A, et al. Synapsins:mosaics of shared and individual domains in a family of synaptic vesicle phos-phoproteins. Science 1989;245:1474–80.

[10] Mandell JW, Townes-Anderson E, Czernik AJ, Cameron R, Greengard P,De Camilli P. Synapsins in the vertebrate retina: absence from ribbonsynapses and heterogeneous distribution among conventional synapses. Neu-ron 1990;5:19–33.

[11] Mandell JW, Czernik AJ, De Camilli P, Greengard P, Townes-Anderson E. Differ-ential expression of synapsins I and II among rat retinal synapses. J Neurosci1992;12:1736–49.

[12] Stone LM, Browning MD, Finger TE. Differential distribution of the synapsins inthe rat olfactory bulb. J Neurosci 1994;14:301–9.

[13] Kielland A, Erisir A, Walaas SI, Heggelund P. Synapsin utilization differsamong functional classes of synapses on thalamocortical cells. J Neurosci2006;26:5786–93.

[14] Bogen IL, Boulland J, Mariussen E, Wright MS, Fonnum F, Kao H, et al. Absence ofsynapsin I and II is accompanied by decreases in vesicular transport of specificneurotransmitters. J Neurochem 2006;96:1458–66.

[15] Bragina L, Candiracci C, Barbaresi P, Giovedì S, Benfenati F, Conti F. Hetero-geneity of glutamatergic and GABAergic release machinery in cerebral cortex.Neuroscience 2007;146:1829–40.

[16] Micheva KD, Busse B, Weiler NC, O’Rourke N, Smith SJ. Single-synapse analysisof a diverse synapse population: proteomic imaging methods and markers.Neuron 2010;68:639–53.

[17] Bragina L, Giovedì S, Barbaresi P, Benfenati F, Conti F. Heterogeneity of glu-tamatergic and GABAergic release machinery in cerebral cortex: analysisof synaptogyrin, vesicle-associated membrane protein, and syntaxin. Neuro-science 2010;165:934–43.

[18] Grønborg M, Pavlos NJ, Brunk I, Chua JJE, Münster-Wandowski A, Riedel D,et al. Quantitative comparison of glutamatergic and GABAergic synaptic vesi-cles unveils selectivity for few proteins including MAL2, a novel synaptic vesicleprotein. J Neurosci 2010;30:2–12.

[19] Pieribone VA, Porton B, Rendon B, Feng J, Greengard P, Kao H-. Expression ofsynapsin III in nerve terminals and neurogenic regions of the adult brain. J CompNeurol 2002;454:105–14.

[20] Kao H, Li P, Chao HM, Janoschka S, Pham K, Feng J, et al. Early involvement ofsynapsin III in neural progenitor cell development in the adult hippocampus. JComp Neurol 2008;507:1860–70.

[21] Lohmann SM, Ueda T, Greengard P. Ontogeny of synaptic phosphoproteins inbrain. Proc Natl Acad Sci USA 1978;75:4037–41.

[22] Mason CA. Axon development in mouse cerebellum: embryonic axon formsand expression of synapsin I. Neuroscience 1986;19:1319–33.

[23] Haas CA, DeGennaro LJ. Multiple synapsin I messenger RNAs are differ-entially regulated during neuronal development. J Cell Biol 1988;106:195–203.

[24] Melloni Jr RH, DeGennaro LJ. Temporal onset of synapsin I gene expressioncoincides with neuronal differentiation during the development of the nervoussystem. J Comp Neurol 1994;342:449–62.

[25] Bogen IL, Jensen V, Hvalby O, Walaas SI. Synapsin-dependent development ofglutamatergic synaptic vesicles and presynaptic plasticity in postnatal mousebrain. Neuroscience 2009;158:231–41.

[26] Ferreira A, Kao H, Feng J, Rapoport M, Greengard P. Synapsin III developmen-tal expression, subcellular localization, and role in axon formation. J Neurosci2000;20:3736–44.

[27] Cesca F, Baldelli P, Valtorta F, Benfenati F. The synapsins: key actors of synapsefunction and plasticity. Prog Neurobiol 2010;91:313–48.

[28] Fornasiero EF, Bonanomi D, Benfenati F, Valtorta F. The role of synapsins inneuronal development. Cell Mol Life Sci 2010;67:1383–96.

[29] Betz WJ, Bewick GS. Optical analysis of synaptic vesicle recycling at the frogneuromuscular junction. Science 1992;255:200–3.

[30] Miesenbòck G, De Angelis DA, Rothman JE. Visualizing secretion andsynaptic transmission with pH-sensitive green fluorescent proteins. Nature1998;394:192–5.

[31] Sankaranarayanan S, De Angelis D, Rothman JE, Ryan TA. The use of pHluo-rins for optical measurements of presynaptic activity. J Biophys 2000;79:2199–208.

[32] Ryan TA, Li L, Chin L, Greengard P, Smith SJ. Synaptic vesicle recycling insynapsin I knock-out mice. J Cell Biol 1996;134:1219–27.

[33] Feng J, Chi P, Blanpied TA, Xu Y, Magarinos AM, Ferreira A, et al. Regulation ofneurotransmitter release by synapsin III. J Neurosci 2002;22:4372–80.

[34] Chi P, Greengard P, Ryan TA. Synapsin dispersion and reclustering during synap-tic activity. Nat Neurosci 2001;4:1187–93.

[35] Chi P, Greengard P, Ryan TA. Synaptic vesicle mobilization is regulated by dis-tinct synapsin I phosphorylation pathways at different frequencies. Neuron2003;38:69–78.

[36] Menegon A, Bonanomi D, Albertinazzi C, Lotti F, Ferrari G, Kao H, et al. Proteinkinase A-mediated synapsin I phosphorylation is a central modulator of Ca2+-dependent synaptic activity. J Neurosci 2006;26:11670–81.

[37] Mozhayeva MG, Sara Y, Liu X, Kavalali ET. Development of vesicle pools duringmaturation of hippocampal synapses. J Neurosci 2002;22:654–65.

[38] Gitler D, Cheng Q, Greengard P, Augustine GJ. Synapsin IIa controls the reservepool of glutamatergic synaptic vesicles. J Neurosci 2008;28:10835–43.

[39] Leal-Ortiz S, Waites CL, Terry-Lorenzo R, Zamorano P, Gundelfinger ED, Gar-ner CC. Piccolo modulation of Synapsin1a dynamics regulates synaptic vesicleexocytosis. J Cell Biol 2008;181:831–46.

[40] Li J, Han YR, Plummer MR, Herrup K. Cytoplasmic ATM in neurons modulatessynaptic function. Curr Biol 2009;19:2091–6.

[41] Granseth B, Odermatt B, Royle SJ, Lagnado L. Clathrin-mediated endocytosis isthe dominant mechanism of vesicle retrieval at hippocampal synapses. Neuron2006;51:773–86.

[42] Messa M, Congia S, Defranchi E, Valtorta F, Fassio A, Onofri F, et al. Tyrosinephosphorylation of synapsin I by src regulates synaptic-vesicle trafficking. JCell Sci 2010;123:2256–65.

[43] Kim SH, Ryan TA. CDK5 serves as a major control point in neurotransmitterrelease. Neuron 2010;67(September (9)):797–809.

[44] Moulder KL, Jiang X, Taylor AA, Shin W, Gillis KD, Mennerick S. Vesiclepool heterogeneity at hippocampal glutamate and GABA synapses. J Neurosci2007;27:9846–54.

[45] Takei Y, Harada A, Takeda S, Kobayashi K, Terada S, Noda T, et al. Synapsin Ideficiency results in the structural change in the presynaptic terminals in themurine nervous system. J Cell Biol 1995;131:1789–800.

[46] Gitler D, Takagishi Y, Feng J, Ren Y, Rodriguiz RM, Wetsel WC, et al. Differentpresynaptic roles of synapsins at excitatory and inhibitory synapses. J Neurosci2004;24:11368–80.

[47] Siksou L, Rostaing P, Lechaire J, Boudier T, Ohtsuka T, Fejtová A, et al.Three-dimensional architecture of presynaptic terminal cytomatrix. J Neurosci2007;27:6868–77.

[48] Benfenati F, Valtorta F, Chieregatti E, Greengard P. Interaction of free and synap-tic vesicle-bound synapsin I with F-actin. Neuron 1992;8:377–86.

[49] Benfenati F, Valtorta F, Rossi MC, Onofri F, Sihra T, Greengard P. Interac-tions of synapsin I with phospholipids: possible role in synaptic vesicleclustering and in the maintenance of bilayer structures. J Cell Biol 1993;123:1845–55.

[50] Ceccaldi P, Grohovaz F, Benfenati F, Chieregatti E, Greengard P, Valtorta F.Dephosphorylated synapsin I anchors synaptic vesicles to actin cytoskeleton:an analysis by videomicroscopy. J Cell Biol 1995;128:905–12.

[51] Jensen V, Walaas SI, Hilfiker S, Ruiz A, Hvalby Ø. A delayed response enhance-ment during hippocampal presynaptic plasticity in mice. J Physiol (Lond)2007;583:129–43.

[52] Hvalby Ø, Jensen V, Kao H, Walaas SI. Synapsin-regulated synaptic transmissionfrom readily releasable synaptic vesicles in excitatory hippocampal synapsesin mice. J Physiol (Lond) 2006;571:75–82.

[53] Bogen IL, Haug KH, Roberg B, Fonnum F, Walaas SI. The importance ofsynapsin I and II for neurotransmitter levels and vesicular storage in cholin-ergic, glutamatergic and GABAergic nerve terminals. Neurochem Int 2009;55:13–21.

[54] Terada S, Tsujimoto T, Takei Y, Takahashi T, Hirokawa N. Impairment ofinhibitory synaptic transmission in mice lacking synapsin I. J Cell Biol1999;145:1039–48.

[55] Baldelli P, Fassio A, Valtorta F, Benfenati F. Lack of synapsin I reduces the readilyreleasable pool of synaptic vesicles at central inhibitory synapses. J Neurosci2007;27:13520–31.

[56] Chiappalone M, Casagrande S, Tedesco M, Valtorta F, Baldelli P, Martinoia S,et al. Opposite changes in glutamatergic and GABAergic transmission underlie

A. Fassio et al. / Seminars in Cell & Developmental Biology 22 (2011) 408– 415 415

the diffuse hyperexcitability of synapsin I-deficient cortical networks. CerebCortex 2009;19:1422–39.

[57] Corradi A, Zanardi A, Giacomini C, Onofri F, Valtorta F, Zoli M, et al. Synapsin-I- and synapsin-II-null mice display an increased age-dependent cognitiveimpairment. J Cell Sci 2008;121:3042–51.

[58] Boido D, Farisello P, Cesca F, Ferrea E, Valtorta F, Benfenati F, et al.Cortico-hippocampal hyperexcitability in synapsin I/II/III knockout mice:age-dependency and response to the antiepileptic drug levetiracetam. Neu-roscience 2010;171:268–83.