Homeostatic Synaptic Plasticity: Local and Global ... · Homeostasis of neuronal ﬁring through...

Homeostatic Synaptic Plasticity: Localand Global Mechanisms for StabilizingNeuronal Function

Gina Turrigiano

Department of Biology and Center for Behavioral Genomics, Brandeis University, Waltham,Massachusetts 02493

Correspondence: [email protected]

Neural circuits must maintain stable function in the face of many plastic challenges, includ-ing changes in synapse number and strength, during learning and development. Recent workhas shown that these destabilizing influences are counterbalanced by homeostatic plasticitymechanisms that act to stabilize neuronal and circuit activity. One such mechanism is syn-aptic scaling, which allows neurons to detect changes in their own firing rates through a set ofcalcium-dependent sensors that then regulate receptor trafficking to increase or decrease theaccumulation of glutamate receptors at synaptic sites. Additional homeostatic mechanismsmay allow local changes in synaptic activation to generate local synaptic adaptations, andnetwork-wide changes in activity to generate network-wide adjustments in the balancebetween excitation and inhibition. The signaling pathways underlying these various formsof homeostatic plasticity are currently under intense scrutiny, and although dozens of mol-ecular pathways have now been implicated in homeostatic plasticity, a clear picture ofhow homeostatic feedback is structured at the molecular level has not yet emerged. On afunctional level, neuronal networks likely use this complex set of regulatory mechanismsto achieve homeostasis over a wide range of temporal and spatial scales.

More than 50 years ago, Walter Cannonmarveled that “somehow the unstable

stuff of which we are composed has learnedthe trick of maintaining stability” (Cannon1932). Along with Claude Bernard, Cannonhad realized that complex physiological systemsare built in a way that promotes stability, or“homeostasis,” of key physiological parameterssuch as body temperature and blood glucoselevels. Recently it has become clear that neu-ronal activity is itself a key physiological param-

eter that is subject to homeostatic regulation.During development billions of neurons wirethemselves up into complex networks and man-age to reach a state where they can generate—and then maintain—stable activity patternsthroughout the life of an organism. What makesthis so extraordinary is that these circuits are notstatic, but are constantly undergoing modifica-tions to allow organisms to store informationand adapt their behavior to a changing environ-ment. Somehow the forces that generate flexi-

Editors: Morgan Sheng, Bernardo Sabatini, and Thomas Sudhof

Additional Perspectives on The Synapse available at www.cshperspectives.org

Copyright # 2012 Cold Spring Harbor Laboratory Press; all rights reserved; doi: 10.1101/cshperspect.a005736

Cite this article as Cold Spring Harb Perspect Biol 2012;4:a005736

1

on March 10, 2020 - Published by Cold Spring Harbor Laboratory Press http://cshperspectives.cshlp.org/Downloaded from

http://cshperspectives.cshlp.org/

bility and those that generate stability are able tocoexist without interfering with each other, andtogether enable the remarkable ability of organ-isms to generate behaviors that are both coher-ent and adaptable.

Neural circuits are subject to many forcesthat work to destabilize their activity. For exam-ple, synapse-specific correlation-based plasti-city mechanisms such as long-term potentia-tion (LTP) and depression (LTD) are widelythought to contribute to learning and informa-tion storage, but theoreticians have long appre-ciated that they generate a powerful destabiliz-ing force on network function (Miller andMacKay 1994; Abbott and Nelson 2000). Thisis because when synapses undergo LTP theyare better able to depolarize the postsynapticneurons, which will increase the probabilitythat they will undergo further LTP—leading tounconstrained synaptic strengthening. A relatedproblem is that as correlated activity drivesstrengthening of specific synapses, and the post-synaptic neuron is driven more strongly, syn-apses that initially were only poorly correlatedwith postsynaptic firing will be better able tofire the postsynaptic neuron, so they too can be-come strengthened. Thus, without forces thatprevent the excitability of the postsynapticneuron from changing in response to correla-tion-based plasticity mechanisms, their specif-icity breaks down and information can nolonger be effectively stored through differencesin synaptic strengths. Many forms of correla-tion-based plasticity have now been describedbiologically, and each is likely to introduce itsown unique destabilizing influences on neuralcircuit function (Abbott and Nelson 2000).

There has been a recent explosion of workinto the biological solutions that neurons andnetworks use to stabilize activity, and collec-tively these mechanisms have been termed “ho-meostatic plasticity” (Marder and Prinz 2003;Turrigiano and Nelson 2004; Davis 2006; Turri-giano 2008; Pozo and Goda 2010). To be consid-ered truly homeostatic, a plasticity mechanismshould regulate a key parameter (such as averageneuronal firing rate) around some set-pointvalue (Fig. 1A). To accomplish this feat, neuronsmust sense some aspect of “activity,” generate an

error signal when this deviates from a set point,then use this error signal to change excitabilityin the correct direction to move activity back to-ward this set point. Here we will define a ho-meostatic form of plasticity as one that acts tostabilize the activity of a neuron or neuronal cir-cuit around some set-point value; some mech-anisms might be adaptive (in that they tend tooppose instability) without being strictly ho-meostatic. Currently a number of phenomenahave been described that may contribute to sta-bilization of neuronal activity, including theactivity-dependent regulation of intrinsic neu-ronal firing (Marder and Prinz 2003; Zhangand Linden 2003); pre- and postsynaptic formsof excitatory synaptic plasticity, such as synapticscaling, that adjust all of a neuron’s excitatorysynapses up or down in the right direction tostabilize firing (Turrigiano and Nelson 2004;Davis 2006); the balancing of excitation and in-hibition within neuronal networks (Maffei et al.2004; Gonzalez-Islas and Wenner 2006); com-pensatory changes in synapse number (Kirovet al. 1999; Wierenga et al. 2006); metaplasticmechanisms that adjust the relative ease of in-ducing LTP and LTD (Bienenstock et al. 1982);and homeostatic regulation of intrinsic excit-ability (Marder and Goaillard 2006; Turrigiano2011). Although all of these phenomena aretheoretically capable of serving a homeostaticfunction, for many of them such a functionhas not been directly shown. In this collectionon the synapse, I will focus on synaptic mecha-nisms that are likely to contribute to networkhomeostasis.

HOMEOSTATIC PLASTICITY AT THENEUROMUSCULAR JUNCTION

The stability problem at the neuromuscularjunction (NMJ) is fairly simple: As muscle fi-bers grow, motor neurons must remain capableof bringing them over threshold for contraction.This matching of motoneuron and muscle fiberproperties is accomplished through an activeprocess that requires signaling between moto-neuron and muscle, and can keep the gain of neu-romuscular coupling remarkably constant. Someof the first reports of homeostatic compensation

G. Turrigiano

2 Cite this article as Cold Spring Harb Perspect Biol 2012;4:a005736



of excitability in the nervous system came fromthe denervation supersensitivity literature (Bergand Hall, 1975; Sharpless 1975), where it wasobserved that loss of synaptic innervation initi-ated an increase in muscle excitability throughchanges in postsynaptic receptor number andlocalization. More recently work at the mam-malian and invertebrate NMJ has shown thatperturbations in postsynaptic function canlead to compensatory changes in presynaptic re-lease, and vice versa (Davis and Bezprozvanny2001).

In Drosophila the ease of genetic manipula-tions has made it possible to perturb either thepresynaptic or postsynaptic side of the synapseand observe exquisitely precise compensation,so that the gain of transmission remains es-sentially constant. For example, reductions inglutamate receptor function, or chronic hyper-

polarization of the muscle, lead to compen-satory increases in transmitter release that re-store evoked transmission to control levels(Paradis et al. 2001). The mechanisms underly-ing this presynaptic form of compensation haverecently been worked out in some detail. Com-pensation can occur very quickly when inducedby pharmacological blockade of postsynapticglutamate receptors and occurs in responseto subtle changes in the amplitude of excitatorypostsynaptic potentials. A reduction in postsyn-aptic depolarization generates an as-yet uniden-tified retrograde signal that leads to enhancedtransmitter release (Frank et al. 2006). Other el-ements of this signaling pathway include a pre-synaptic Eph receptor, the Eph interacting pro-tein Exn, and activation of the Rho GTPaseCdc42. This signaling pathway converges ontothe presynaptic calcium channel Cav2.1 to

Firing rate homeostasisA

Synaptic scalingB

Lower firing

Raise firing

Presynapticterminal Postsynaptic

terminalPotentiation Scaling

Target activityTarget activity High activity

Figure 1. Homeostasis of neuronal firing through homeostatic synaptic plasticity. (A) Cartoon illustration of thephenomenon of firing rate homeostasis in dissociated neocortical networks; perturbing firing in either directionresults in the homeostatic regulation of synaptic and intrinsic properties so that baseline firing rates are restored.(B) One mechanism contributing to the firing rate homeostasis illustrated in A is synaptic scaling. When activityis perturbed (illustrated here as the potentiation of some inputs through Hebbian mechanisms) this triggerssynaptic scaling, which produces a proportional reduction in strength at all synapses of the right magnitudeto return firing to baseline levels. Note that, because this mechanism scales synaptic strength up or down propor-tionally, the relative difference in synaptic strengths induced by Hebbian mechanisms is preserved.

Homeostatic Synaptic Plasticity

Cite this article as Cold Spring Harb Perspect Biol 2012;4:a005736 3



enhance presynaptic calcium influx and thusneurotransmitter release (Frank et al. 2009). Afunctionally similar mechanism coupling post-synaptic muscle activation with presynaptic reg-ulation of neurotransmitter release has recentlybeen described at the Caenorhabditis elegansNMJ (Simon et al. 2008). It should be notedthat there is evidence at the Drosophila NMJfor additional parallel signaling pathways dur-ing presynaptic compensation (Frank et al.2006; Goold and Davis 2007; Frank et al.2009), as well as mechanisms that can inducepostsynaptic compensatory changes in quantalamplitude (Davis and Goodman 1998). Thiscomplexity in compensatory mechanisms andsignaling pathways is a theme we shall returnto when we discuss homeostatic plasticity atcentral synapses.

HOMEOSTATIC PLASTICITY AT CENTRALMAMMALIAN SYNAPSES

There is compelling evidence from experimentsin neuronal cultures that central neurons areable to maintain average firing rates around ahomeostatic set point. When cortical or hippo-campal neurons are induced to fire more thannormal, over many hours firing returns to base-line levels, and similarly, if neuronal firing isreduced over time, neurons compensate andagain firing is restored (Fig. 1A) (Turrigianoet al. 1998; Burrone et al. 2002). Central neuronsare embedded in complex networks composedof many cell types, including both excitatoryand inhibitory neurons, and small changes inthe balance between excitation and inhibitioncan have major effects on ongoing activity (Nel-son and Turrigiano 2008). The ability to com-pensate for external perturbations and maintainstable firing is thus not trivial, and it is not sur-prising that a rich variety of homeostatic mech-anisms operating over various temporal andspatial scales are likely to contribute to thisprocess. For example, there is evidence forboth “global” mechanisms that operate on allof a neuron’s synapses (Turrigiano 2008), and“local” mechanisms that act on individual orsmall groups of synapses (Yu and Goda 2009).Similarly, some forms of homeostatic plasticity

occur through presynaptic and others throughpostsynaptic changes in function (Davis andBezprozvanny 2001). Which of these mecha-nisms are engaged will likely depend on howactivity is modulated, as well as other factorssuch as cell type and developmental stage.Below we will review the current state of under-standing of the signaling pathways and expres-sion mechanisms of several forms of homeo-static synaptic plasticity at central mammaliansynapses.

CELL-AUTONOMOUS, GLOBAL SYNAPTICSCALING OF EXCITATORY SYNAPSES

Currently the best understood form of homeo-static plasticity at central excitatory synapses issynaptic scaling. Synaptic scaling was first iden-tified in cultured neocortical neurons, where itwas observed that perturbing network activitygenerated compensatory changes in synapticstrength that were in the right direction to re-store average firing rates to baseline values(Turrigiano et al. 1998). Pharmacological ma-nipulations of activity are able to induce bidir-ectional compensatory changes in the unitstrength of glutamatergic synapses, which canbe measured by recording miniature excitatorypostsynaptic currents (mEPSCs, or “minis”).Minis represent the postsynaptic response to re-lease of individual vesicles of neurotransmitters;by measuring minis arising from many synapsesonto the same neuron, it was observed thatmodulating network activity induced uniformincreases or decreases in the entire mini ampli-tude distribution, in effect scaling postsynapticstrength up or down (Turrigiano et al. 1998;Desai et al. 2002; Gainey et al. 2009). Interest-ingly, synaptic scaling protocols homeostati-cally regulate both the NMDA and the AMPAcomponent of glutamatergic synaptic currents,and the change in AMPA and NMDA currentsare proportional at individual synapses (Wattet al. 2000; Perez-Otano and Ehlers 2005). Thesechanges in mini amplitude translate into changesin the amplitude of evoked transmission, withlittle or no change in short-term synapticdynamics (Watt et al. 2000; Maffei et al. 2004;Wierenga et al. 2005). Such a postsynaptic

G. Turrigiano




scaling process has the attractive property ofallowing neurons to stabilize activity withoutchanging the relative strength of synaptic inputs,thus avoiding disrupting information storage orprocessing mechanisms that rely on differencesin synaptic weights (Fig. 1B). Synaptic scalinghas now been shown in a variety of central neu-rons both in vitro and in vivo, including neo-cortical and hippocampal pyramidal neuronsand spinal neurons (O’Brien et al. 1998; Turri-giano et al. 1998; Desai et al. 2002; Stellwagenand Malenka 2006; Goel and Lee 2007; Kim andTsien 2008; Knogler et al. 2010). Currently, mostof the mechanistic work on synaptic scaling hasinvolved neocortical or hippocampal pyramidalneurons in dissociated culture; whether themechanisms that underlie synaptic scaling invivo and in other brain regions are similar or dis-tinct to those for cultured cortical neurons re-mains an open question.

Perturbations in network activity could besensed by individual neurons as changes intheir own firing, local changes in receptor acti-vation, or changes in release of secreted factors.Although some forms of homeostatic plasticityappear to be triggered by local signaling orsignaling through secreted factors (these willbe reviewed in turn below), there is strong evi-dence that synaptic scaling is a cell-autonomousprocess induced by changes in a neuron’s ownfiring. Selectively blocking firing by microper-fusion of TTX to the soma of individual neu-rons scales up synaptic strengths to the samedegree as blockade of network activity, whereaslocal block of synaptic transmission does notinduce a local enhancement of receptor accumu-lation (Ibata et al. 2008). Interestingly, chronichyperpolarization of an entire neuron by expres-sion of an inwardly rectifying K channel (Kir)does not induce postsynaptic scaling (Burroneet al. 2002; Hou et al. 2008), suggesting thatchronic global hyperpolarization of neuronaldendrites and somata overrides the signals thatinduce scaling up in response to a drop in firing.

Selectively blocking postsynaptic firing inneocortical neurons scales synapses up througha process that requires a drop in somatic cal-cium influx, reduced activation of CaMKIV,and transcription (Ibata et al. 2008). This sig-

naling pathway then leads to enhanced accu-mulation of AMPA-type glutamate receptors(AMPAR) in the postsynaptic membrane at allexcitatory synapses, thus scaling up mini ampli-tude and enhancing evoked transmission. Thisglobal enhancement of AMPAR abundance inresponse to activity blockade requires sequenceson the carboxyl terminus of the GluA2 subunitof the AMPAR (Gainey et al. 2009), and themajority of studies that have blocked activitywith tetrodotoxin (TTX) or AMPAR antago-nists have observed a coordinated increase inGluA1 and GluA2 (O’Brien et al. 1998; Wier-enga et al. 2005; Cingolani et al. 2008; Sunand Wolf 2009; Anggono et al. 2011; but seeThiagarajan et al. 2005). Additionally, there isevidence that the neurotrophin brain-derivedneurotrophic factor (BDNF) (Rutherford et al.1998), the immediate early gene Arc (Shepherdet al. 2006), the cytokine TNFa (Stellwagen andMalenka 2006; Steinmetz and Turrigiano,2010), the immune molecule MHC1 (Goddardet al. 2007), b3 integrins (Cingolani et al. 2008),the AMPAR binding protein PICK1 (Anggonoet al. 2011), and the scaffold proteins PSD-95and PSD-93 (Sun and Turrigiano 2011) can allcontribute to, or are essential for, synaptic scal-ing in dissociated cultures. Several of these mol-ecules are known to regulate AMPA receptortrafficking; for example, Arc interacts with theendocytic machinery that removes AMPARfrom the membrane (Chowdhury et al. 2006),TNFa directly increases synaptic AMPAR accu-mulation (Beattie et al. 2002; Stellwagen et al.2005), b3 integrins regulate AMPAR surface ex-pression (Cingolani et al. 2008), PICK1 regu-lates the pool size of intracellular receptors(Anggono et al. 2011), and PSD-95 stabilizesAMPAR at synaptic sites (Bats et al. 2007); buthow these various players cooperate to accom-plish the homeostatic regulation of AMPARabundance remains unclear.

Like scaling up, scaling down in responseto elevated network activity is a cell-autono-mous function of postsynaptic firing, and in-volves enhanced calcium influx (Fig. 2), genetranscription, the CaMKK/CaMKIV signalingpathway, and targets the GluA2 subunit (Gooldand Nicoll 2010). Another calcium-dependent





pathway implicated in scaling down is the polo-like kinase 2 (Plk2)-CDK5 signaling pathway(Seeburg et al. 2008). Calcium influx activatesPlk2, which when primed by CDK5 can bindto the scaffold protein Spar and trigger its deg-radation, and this pathway is necessary for thereduction in synaptic AMPAR accumulation in-duced by elevated activity (Seeburg and Sheng2008; Seeburg et al. 2008). Plk2 can also reducesurface AMPAR number through a kinase-independent association with NSF (Evers et al.2011), suggesting Plk2 may operate throughparallel pathways to reduce synaptic strengthin response to elevated activity. Two additionalcalcium-dependent pathways have recentlybeen identified. One involves the immediateearly gene Homer1a, which is induced in acalcium-dependent manner by enhanced activ-ity and is required for scaling down through apathway that requires agonist-independent reg-ulation of mGluRs and reduced tyrosine phos-phorylation of GluA2 (Hu et al. 2010). Thesecond involves the activation of Eph4A by

elevated activity; Eph4A is also necessary forscaling down, and is thought to act by regulatingthe ubiquitin pathway to control AMPAR deg-radation (Fu et al. 2011). Thus, raising networkfiring activates a slew of calcium-dependent sig-naling pathways, each with elements that arerequired for the expression of scaling down.One explanation for why so many apparentlyindependent calcium-activated signaling path-ways are all necessary for the expression ofscaling down is that these signals converge atsome point in the scaling pathway to coordin-ately regulate a critical step in synaptic AMPARreduction. Alternatively, there may be multiplesteps in the scaling pathway that must all beregulated to induce scaling, each of which is tar-geted by a separate calcium-dependent signal-ing pathway.

There is conflicting evidence as to whetherscaling up and down involve reciprocal mod-ulation of the same signaling pathways. Al-though scaling up and down have in commonreciprocal modulation of somatic calcium influx

Scalingup

Scalingup

Scalingup

[Ca]i

[Ca]i [Ca]i

[Ca]i

Scalingdown

Scalingdown

Scalingdown

A

B C

Figure 2. Calcium-dependent pathways regulate both scaling up and scaling down. (A) At a particular averagelevel of somatic calcium influx, scaling up and scaling down will balance each other, and the resulting synapticequilibrium will help determine the firing rate set point of the neuron. (B) If activity decreases (owing to sensorydeprivation, learning-induced LTD, or other factors) then average somatic calcium will also decrease; this willenhance scaling up and reduce scaling down and restore firing to baseline. (C) Conversely, if firing increases andaverage somatic calcium increases, this will enhance scaling down and reduce scaling up, again restoring firing tobaseline.

G. Turrigiano




(Fig. 2), and both involve changes in signalingthrough the CaMKK/CaMKIV pathways (Ibataet al. 2008; Goold and Nicoll 2010), severalother signaling and trafficking elements knownto be important or essential for scaling upare dispensable for scaling down, includingBDNF, TNFa, Arc, and PICK1 (Rutherfordet al. 1998; Leslie et al. 2001; Shepherd et al.2006; Stellwagen and Malenka 2006). Further,whereas scaling up is completely dependenton the scaffold protein PSD-95, scaling downcan be supported by either PSD-95 or PSD-93(Sun and Turrigiano 2011). Given the differ-ential dependence of scaling up and down onvarious aspects of the synaptic regulatorymachinery, it seems likely that the signalingpathways involved diverge shortly after thecommon trigger of altered calcium influx.This suggests that a drop in somatic calcium in-flux will simultaneously enhance signaling inthe pathways that scale synapses up and reducesignaling in the pathways that scale synapsesdown, and vice versa when firing rises. Thehomeostatic control of quantal amplitude isthus likely to work in a push-pull manner, withopposing signaling pathways acting to scale syn-apses up or down. One interesting implicationof this model is that the firing rate set point isan emergent property that represents the stablepoint of all the opposing signaling pathwaysthat operate on synaptic strength (Turrigiano2008).

As the above discussion should have madeabundantly clear, the molecular events thatlead from cell-autonomous changes in firingto changes in synaptic AMPAR abundance areincompletely understood. As for homeostaticregulation at the NMJ, it seems that a numberof signaling pathways are activated by synapticscaling protocols, and dozens of moleculeshave been identified that contribute in someway to the ability of synapses to adaptivelyenhance or reduce receptor accumulation. Al-though some of these players are undoubtedlyintegral signaling elements for scaling, otherswill instead turn out to be permissive, eitherbecause they act constitutively to maintain syn-apses in a plastic state (for example, TNFa)(Steinmetz and Turrigiano 2010), or because

they are part of the constitutive trafficking path-ways that maintain synaptic receptor abun-dance.

Interestingly, the dependence of synapticscaling on GluA2 differentiates it from otherforms of synaptic enhancement, such as localhomeostatic plasticity induced by concur-rent TTX and N-methyl-D-aspartate receptor(NMDAR) block (see below), and many formsof LTP that require regulatory sequences onthe GluA1 rather than the GluA2 subunit (Ma-lenka and Bear 2004). Several recent studieshave now reinforced the idea that the traffick-ing mechanisms that underlie enhanced or re-duced AMPAR accumulation during homeo-static plasticity are fundamentally differentfrom those that underlie Hebbian forms ofplasticity such as LTP and LTD. For example,whereas LTD depends on the SH3-GK domainsof PSD-95 (Xu et al. 2008), scaling downdepends on the PDZ1/2 domains (which in-teract with transmembrane AMPA receptorregulatory proteins [TARPs]) (Sun and Turri-giano 2011). Further, enhancement of synapticstrength during LTP and scaling up are differen-tially dependent on PSD-95; KD does not blockLTP but does block scaling up, whereas overex-pression occludes LTP but not scaling up (Steinet al. 2003; Ehrlich and Malinow 2004; Ehrlichet al. 2007; Sun and Turrigiano 2011). Similarly,whereas several forms of LTD depend on PICK1(which binds to GluA2-containing AMPAR)(Steinberg et al. 2006; Terashima et al. 2008;Volk et al. 2010), scaling down is unaffected byloss of PICK1 (Anggono et al. 2011). Thus,cell-autonomously induced synaptic scaling isfundamentally different from LTP and LTD: itoperates over a longer temporal scale (hours),a wider spatial scale (global), and utilizes dis-tinct trafficking steps and molecular machineryto enhance or reduce the synaptic accumulationof GluA2-containing AMPAR at most if not allsynapses onto a neuron.

It is widely agreed that synaptic scalinginvolves postsynaptic changes in receptor accu-mulation (Turrigiano 2008), but under somecircumstances additional presynaptic changescan also be recruited (Davis and Bezprozva-nny 2001; Rich and Wenner 2007; Thiagarajan





et al. 2007). For example, there is evidence thatglutamate transporter (VGlut) expression in thepresynaptic terminal can be enhanced by activ-ity blockade (De Gois et al. 2005; Erickson et al.2006), raising the possibility that in addition toadding more postsynaptic receptors, neuronsmight also package more neurotransmittersinto each presynaptic vesicle; these two proc-esses would then collaborate to scale up miniamplitude. It is worth noting, however, thatpostsynaptic knockdown of GluA2 is able tocompletely block synaptic scaling (Gaineyet al. 2009), suggesting that if there is a pre-synaptic change in transmitter packaging it isnot sufficient to enhance mini amplitude onits own. Thus the physiological significanceof changes in presynaptic VGlut remain to bedetermined.

Interestingly, the same manipulation (activ-ity blockade using TTX or AMPAR antagonists)that only affects mini amplitude in young cul-tures (Turrigiano et al. 1998; Wierenga et al.2005) can also induce changes in mini fre-quency and in presynaptic release probabilityin older neuronal cultures (Wierenga et al.2006), through a process that involves post-synaptic hyperpolarization and a drop in cal-cium influx (Burrone et al. 2002; Thiagarajanet al. 2005); why this transition occurs withtime in culture is not currently known. Like syn-aptic scaling, this presynaptic form of plasticitycan also be induced in a cell-autonomous man-ner by a reduction in postsynaptic depolariza-tion (Burrone et al. 2002). The locus of changeduring homeostatic plasticity has importantconsequences for circuit function, because(for example) presynaptic changes in releaseprobability will strongly affect short-term plas-ticity and thus the dynamics of informationtransfer across synapses, whereas a postsynapticchange in receptor accumulation can scale post-synaptic responsiveness without affecting theshort-term dynamics of synaptic transmission(Abbott and Nelson 2000). Given that the post-synaptic and presynaptic responses to activitydeprivation can occur independently (Wier-enga et al. 2005, 2006), it is likely that they rep-resent distinct forms of plasticity induced bydistinct signaling pathways.

LOCAL AND QUASI-LOCAL FORMS OFHOMEOSTATIC SYNAPTIC PLASTICITYIN CENTRAL NEURONS

In addition to forms of homeostatic plasticitysuch as synaptic scaling that are induced in aglobal manner as a function of postsynaptic fir-ing, there is also evidence that local or quasi-local changes in synaptic signaling can inducehomeostatic changes in synaptic strength. Atruly local form of plasticity would be inducedat an individual synapse as a function of changesin presynaptic release and/or postsynaptic re-ceptor activation at an individual synaptic site.Several studies have now looked at the effectsof local changes in signaling on postsynapticreceptor accumulation, with mixed results.Lowering presynaptic firing with an inwardlyrectifying Kir channel was observed to selec-tively enhance the accumulation of GluA1 atpostsynaptic sites, suggesting the existence of asynapse-specific form of homeostatic compen-sation to reduced receptor activation (Houet al. 2008). In contrast, three other studiesfailed to observe such a local enhancement ofAMPAR accumulation at inactivated synapses.Local perfusion of a small number of synapticsites with either TTX to block presynapticspikes or DNQX/APV to block postsynapticglutamate receptors failed to enhance post-synaptic AMPAR accumulation (Ibata et al.2008), as did chronic block of release at a subsetof synapses by presynaptic expression of tetanustoxin (Harms and Craig 2005; Ehlers et al.2007). The ability of synapses to undergo post-synaptic homeostatic regulation as a function ofpurely local changes in transmitter release andreceptor activation thus remains controversialand may require either extreme presynaptic hy-perpolarization (such as would be obtainedwith presynaptic expression of Kir) or a situa-tion in which spiking and transmitter releaseare both severely impaired (see discussion belowon the effects of simultaneous application ofTTX and glutamate receptor blockers).

On a theoretical level it is unclear what pur-pose truly local homeostatic regulation wouldserve, as potentiating a synapse through an LTP-like mechanism would then lead to a homeostatic

G. Turrigiano




reduction in strength (and LTD would lead topotentiation), in effect erasing memory storage.On the other hand, theoretical work has sug-gested that quasi-local forms of homeostasisthat act on groups of nearby synapses can servea useful normalization function without se-verely disrupting Hebbian plasticity (Rabino-witch and Segev 2006a,b). Thus, it will beimportant to understand exactly how local“local” homeostatic mechanisms are. Further,it will be important to understand if this mech-anism only operates under extreme conditions(such as when presynaptic firing and transmit-ter release are drastically reduced), or trulyacts homeostatically to keep net synaptic activ-ation constant through graded adjustments inpostsynaptic strength to compensate for changesin presynaptic release.

Interestingly, when global block of firingwith TTX is combined with local glutamatereceptor block, there is a local enhancementof GluA1 accumulation at the blocked synapses(Sutton et al. 2006). This enhancement occursvia a fundamentally different mechanism thanthe global synaptic scaling described above.Enhanced AMPAR accumulation under con-ditions of both action potential and NMDARblock involves local synthesis and insertionof GluA1-containing/GluA2-lacking AMPARinto synapses, so that the synaptic compositionof AMPAR is modified (Ju et al. 2004; Suttonet al. 2006; Aoto et al. 2008). This is in contrastto blocking firing alone, which leads to en-hancement of minis through increased synap-tic accumulation of GluA2-containing AMPAR(see above). It has recently been suggested thatthe enhanced GluA1 synthesis and accumula-tion induced by TTX þ APV is mediatedthrough production of retinoic acid (RA) andactivation of the RA receptor RARa, as sup-pressing RA production or RARa receptorsblocks the effects of TTX þ APV, and RA canstimulate local translation of GluA1 (Aotoet al. 2008; Poon and Chen 2008).

Exactly why and how blocking action po-tential firing enables local glutamate receptorblock to enhance receptor accumulation is notcurrently known. One outstanding question iswhether TTX is acting by blocking presynaptic

or postsynaptic firing; although presynapticblock of action potentials coupled to postsynap-tic block of NMDAR can trigger enhanced den-dritic protein synthesis (Sutton et al. 2007), it isnot known whether this combination is alsosufficient to trigger enhanced GluA1 synthesis,local synaptic insertion, and enhanced synap-tic strength. If presynaptic firing is the relevantsignal during this paradigm then one wouldpredict that presynaptic tetanus toxin, whichblocks both action potentials and neurotrans-mitter release in the presynaptic neuron (Ehlerset al. 2007), should trigger local postsynapticGluA1 accumulation; yet this manipulationdoes not (Harms and Craig 2005; Ehlers et al.2007). Clearly there is more to be done toilluminate the pre- and postsynaptic signalsand signaling pathways that trigger this formof homeostatic plasticity, as well as its functionwithin neuronal circuits.

On the presynaptic side, there is compel-ling evidence that homeostatic modulation ofneurotransmitter release probability (Pr) canhappen at the level of individual dendriticbranches. Enhanced synaptic activity was ob-served to reduce Pr in cultured hippocampalneurons, through a mechanism that was localto particular dendrites (Branco et al. 2008).Synapses onto the same dendrite tended tohave similar Pr, and there was an inverse rela-tionship between synapse number and Pr ontoindividual dendrites, suggesting that this regu-lation happens in a “quasi-local” manner as aresult of the degree of dendritic depolarization.This is consistent with evidence that chronichyperpolarization of entire neurons (includingdendrites) results in enhanced Pr (Burroneet al. 2002), suggesting that Pr can be bidirec-tionally regulated by changes in the amount ofdendritic depolarization. It is currently some-what unclear what the induction requirementsare for this form of local presynaptic plasticity;local synaptic blockade was not sufficient toenhance Pr in one study (Branco et al. 2008),whereas in another local synaptic blockade-induced presynaptic plasticity that was pre-vented by presynaptic firing (Jakawich et al.2010); this requirement for presynaptic firingcould explain why TTX generally does not





induce presynaptic homeostatic plasticity (Tur-rigiano et al. 1998; Wierenga et al. 2005). Thisquasi-local (dendrite-wide) form of presynaptichomeostasis has been suggested to prevent syn-aptic saturation owing to strong depolarizationfrom summed inputs onto a dendrite, by reduc-ing Pr when postsynaptic activation rises toohigh (Branco et al. 2008). If this mechanism op-erates in vivo, it predicts an inverse relationshipbetween Pr and the summed synaptic strengthonto a dendritic branch. In apparent conflictwith this model, a positive correlation betweenPr and synaptic strength at cortical synapseshas been reported (Hardingham et al. 2010;Kay et al. 2011), but because in these studies syn-aptic strength was measured at unitary connec-tions, it remains an open possibility that suchan inverse correlation exists at the level ofsummed dendritic input.

ROLE OF NETWORK ACTIVITY ANDSECRETED FACTORS IN THE INDUCTIONOF HOMEOSTATIC PLASTICITY

In theory homeostatic plasticity mechanismscould also exist at the network level, and operatethrough the activity-dependent release of se-creted factors that act at a number of sites withinthe network to regulate the excitation/inhibi-tion (E/I) balance. BDNF was the first suchsecreted factor suggested to playa role in homeo-static plasticity (Rutherford et al. 1997, 1998).BDNF is thought to be released as a function ofpyramidal neuron activity, and activates Trk re-ceptors on both pyramidal and interneurons(Lu 2003). Blocking BDNF signaling mimics,and exogenous BDNF application prevents, theeffects of activity blockade on excitatory minis,suggesting that in neocortical neurons activity-dependent BDNF release could mediate synapticscaling (Rutherford et al. 1998). However, scal-ing up can also be induced by chronic changesin the level of depolarization in a BDNF-inde-pendent manner (Leslie et al. 2001), and the re-lationship between BDNF-mediated synapticscaling and cell-autonomous synaptic scalingof excitatory inputs (Ibata et al. 2008) remainsunclear. As mentioned previously, these consid-erations suggest that there are several distinct

signaling pathways that are capable of inducingscaling up of synaptic strengths in response toa drop in activity.

Interestingly, postsynaptic BDNF release hasrecently been implicated in the rapid local regu-lation of Pr (Jakawich et al. 2010), and BDNFis clearly important for the development andregulation of inhibitory synaptic transmission(Huang et al. 1999). A number of reports havesuggested that activity-dependent BDNF releaseis critical for scaling of inhibitory synapses ontopyramidal neurons as well as the homeostaticregulation of excitatory synapses onto inhibi-tory neurons (Rutherford et al. 1997, 1998;Copi et al. 2005; Swanwick et al. 2006). Consis-tent with the idea that inhibitory scaling mightresult from release of a secreted factor duringchanges in network activity, scaling of inhibi-tory synapses cannot be induced by hyperpola-rization of individual presynaptic or postsynap-tic neurons (Hartman et al. 2006). The ability ofsecreted BDNF to regulate many synapse typesboth pre- and postsynaptically in a coordinatedhomeostatic manner suggests that activity-de-pendent BDNF release contributes to a formof network-wide homeostatic balancing of exci-tation and inhibition.

Another secreted factor suggested to con-tribute to homeostatic plasticity is TNFa.TNFa is a cytokine that is part of the inflamma-tory response to pathological states (Bessis et al.2007). It was recently proposed that prolongedactivity blockade (with TTX) increases glial re-lease of TNFa, which then acts on neurons toenhance AMPAR insertion and scale up mEPSCamplitude (Stellwagen and Malenka 2006). Thismodel is based on the observations that TNFalevels are elevated by prolonged (48 h) activityblockade, acute application of TNFa (or condi-tioned media from activity-blocked cultures)increases mEPSC amplitude, and scaling up inresponse to prolonged activity blockade is pre-vented by blocking TNFa signaling (Beattieet al. 2002; Stellwagen et al. 2005; Stellwagenand Malenka 2006). Interestingly, the TNFainvolved in scaling up synaptic strengths inresponse to TTX originates from glia ratherthan neurons, because wild-type neurons grownon TNFa2/2 glia did not scale up synaptic

G. Turrigiano




strengths in response to 48 h of TTX treatment,whereas TNFa2/2 neurons grown on wild-typeglia did (Stellwagen and Malenka 2006).

Inconsistent with the above model is the ob-servation that scaling is a gradual and cumula-tive process evident after as little as 4–6 h ofactivity blockade (Sutton et al. 2006; Ibataet al. 2008), whereas TNFa-dependent scalingwas only observed after prolonged activity block(48 h) (Stellwagen and Malenka 2006). Thisraises the possibility that the early phase of scal-ing is not mediated by TNFa, or alternatively,that TNFa is permissive rather than instructivefor scaling. A recent study suggests that the laterinterpretation is the correct one. TNFa signal-ing was not necessary for the induction of early(6 h) scaling, but became essential for main-tenance of scaling during prolonged (.24 h)activity blockade. Further, blocking TNFa sig-naling for 24 h before blocking activity withTTX, prevented both early and late scaling(Steinmetz and Turrigiano 2010). TNFa andTTX also appear to influence AMPAR accu-mulation via distinct mechanisms. Unlike TTXtreatment, acute TNFa application does notscale up mini amplitude and adding TNFa toprescaled synapses actually reduces synapticstrength, indicating that scaling and TNFa areneither additive nor simply occlusive (Stein-metz and Turrigiano 2010). Taken togetherthese studies suggest that glial-derived TNFa re-lease is critical for maintaining synapses in aplastic state in which homeostatic synaptic scal-ing can be expressed.

SPECIFICITY OF SYNAPTIC SCALINGRULES FOR SYNAPSE TYPE

Neural circuits are composed of many excita-tory and inhibitory cell types interconnectedin highly specific ways. Thus, you might expectthat to stabilize the activity of a neural circuityou would need homeostatic plasticity rulesthat are specific for particular classes of synap-ses. In dissociated cortical cultures it has beenshown that excitatory synapses onto excitatorypyramidal neurons are scaled up by activityblockade, whereas excitatory synapses onto GA-BAergic interneurons are not (Rutherford et al.

1998). On the other hand, enhancing networkactivity does increase excitatory transmissiononto GABAergic interneurons (Rutherford et al.1998; Chang et al. 2010), which should promotethe recruitment of additional inhibition whenactivity in a circuit increases. Further, it hasbeen shown both in vitro and in vivo that inhib-itory synapses onto pyramidal neurons are regu-lated in the opposite direction from excitatorysynapses in response to a drop in activity or sen-sory drive (Kilman et al. 2002; Vale and Sanes2002; Maffei et al. 2004; Hartman et al. 2006;Huupponen et al. 2007). Interestingly, both invitro and in vivo studies have documented thathomeostatic regulation of inhibition involvescoordinated changes in postsynaptic strength,synapse number, and likely presynaptic releaseprobability (Kilman et al. 2002; Maffei et al.2004; Hartman et al. 2006). These distinct plas-ticity rules at different classes of excitatory andinhibitory synapse are consistent with an over-all homeostatic shift in the balance betweenexcitation and inhibition, as the net effect ofthe changes induced by activity blockade areto enhance excitation and reduce inhibition,which will act to restore network excitability.

Although in neocortical networks the neteffects of changes in excitation and inhibitionappear to be homeostatic (Rutherford et al.1998; Turrigiano et al. 1998; Maffei et al.2004), different classes of inhibitory synapseare regulated differently by lowered activity.When sensory drive to primary visual cortex islowered in vivo, connections from fast-spikingbasket cells onto layer 4 pyramidal neuronsare reduced in amplitude, whereas connec-tions from another class of interneuron becomesparser but stronger (Maffei et al. 2004). Simi-larly, activity blockade with TTX in neocorticalslice cultures differentially regulates differentclasses of inhibitory synapses (Bartley et al.2008). Further, in hippocampal circuits activityblockade has revealed that under some condi-tions inhibition can change in the same direc-tion as excitation (Echegoyen et al. 2007), butwhether this acts to enhance or oppose stabilityis not entirely clear.

Interestingly, not all excitatory neurons incortical networks express synaptic scaling. In





hippocampal networks CA1 neurons scale syn-apses up in response to activity blockade, where-as CA3 neurons do not (Kim and Tsien 2008).In visual cortex the expression of synaptic scal-ing is developmentally regulated (Desai et al.2002; Maffei et al. 2004, 2006; Maffei and Turri-giano 2008); synaptic scaling is expressed bypyramidal neurons in layer 4 early in postnataldevelopment, but then turns off in layer 4 andbegins to be expressed in layer 2/3 pyramidalneurons around the opening of the classical vis-ual system critical period (Desai et al. 2002;Maffei and Turrigiano 2008), where it persistsinto adulthood (Goel and Lee 2007). An issuethat is not yet resolved is whether postsynapticneurons can differentiate between differenttypes of synapses arising from different sources,and selectively scale one type up or down whileleaving others unaffected. This possibility israised by the finding that the same visual depri-vation paradigm that scales minis up in youngermice produces nonmultiplicative changes inminis in older mice, suggesting that not all syn-apses are affected equally (Goel and Lee 2007);however, this could also reflect the simultane-ous activation of synaptic scaling and Hebbian,synapse-specific mechanisms (see Turrigiano2011 for a discussion of difficulties in interpret-ing scaling data). Interestingly, a recent study onsensory inputs to Xenopus tectum found that, inmultisensory neurons, sensory deprivation ofone modality led to a modality-specific en-hancement in synaptic transmission, suggestingthat synaptic inputs subserving different mo-dalities can undergo independent homeostaticregulation (Deeg and Aizenman 2011). In gen-eral, these studies highlight the point that theforms of homeostatic plasticity present at par-ticular classes of synapse onto particular neuro-nal types will depend critically on their functionwithin the network.

FUNCTIONS OF HOMEOSTATICPLASTICITY IN VIVO

Most work on homeostatic synaptic plasticityhas used in vitro systems to probe functionand uncover molecular mechanisms, but thereis now a growing appreciation that homeostatic

plasticity is a vital aspect of in vivo circuit func-tion at many stages of development. For exam-ple, during embryonic and early postnataldevelopment, homeostatic mechanisms can en-sure that spontaneous activity is present indeveloping spinal circuits (Gonzalez-Islas andWenner 2006; Knogler et al. 2010), where suchactivity is vital for driving proper circuit con-nectivity (Hanson and Landmesser 2004). Sim-ilarly, in visual cortex during the second andthird postnatal weeks when synaptogenesis ishigh, there is an inverse relationship betweenthe frequency and amplitude of mEPSCs ontopyramidal neurons, and this can be preventedby raising animals in the dark (Desai et al.2002). This suggests that as the number of exci-tatory synapses increases (therefore increasingmEPSC frequency) and visual drive increases,synaptic strength is reduced through an activ-ity-dependent homeostatic mechanism. Sucha mechanism could serve the vital developmen-tal function of matching local microcircuit ex-citability to the strength of sensory drive.

Interestingly, in visual cortical microcircuitsthe locus of homeostatic plasticity changes asthe circuit matures. Early in postnatal develop-ment layer 4 (the first input layer to cortex, andthe first layer to mature) responds homeostati-cally to brief (2 days) inactivation of the opticnerve with TTX (Desai et al. 2002; Maffeiet al. 2004), whereas later (at the opening ofthe classical visual system critical period) thishomeostatic response turns off in layer 4 andmigrates to layers 2/3, where it persists intoadulthood (Desai, et al. 2002; Maffei et al.2006; Goel and Lee 2007; Maffei and Turrigiano2008; Goel et al. 2011). Exactly why the locus ofhomeostatic plasticity shifts during develop-ment remains a mystery. One highly speculativepossibility is that later in development (afterthalamocortical inputs are established) homeo-stasis in neocortical layer 4 in response to sen-sory deprivation would only serve to amplifynoise and so becomes maladaptive. In contrast,because neocortical layers 2/3 receive extensivelateral and feedback connections from othercortical areas, homeostasis in layers 2/3 couldserve the useful function in the mature animalof encouraging the takeover of deprived cortical

G. Turrigiano




territory by cortical regions with intact sensorydrive.

A role for homeostatic plasticity has recentlybeen suggested to contribute to ocular domi-nance plasticity in rodent visual cortex (Mrsic-Flogel et al. 2007; Kaneko et al. 2008). When oneeye is deprived of patterned vision during theclassical critical period (using lid suture) thereis a change in the ability of the two eyes to driveneurons within the binocular portion of visualcortex: cortical neurons rapidly (within 2 or 3days) lose responsiveness to the deprived eye,and then more slowly (over 4–6 days) gain re-sponsiveness to the nondeprived eye (Frenkeland Bear 2004; Mrsic-Flogel et al. 2007; Kanekoet al. 2008). Interestingly, binocularly drivenneurons maintained a similar average level ofcombined responsiveness to the two eyes, sug-gesting that the drop in deprived eye responsive-ness was compensated by an increase in nonde-prived eye responsiveness (Mrsic-Flogel et al.2007); a very similar response homeostasis fol-lowing visual deprivation has been reportedat retinotectal synapses (Chandrasekaran et al.2007). Why visual deprivation using lid sutureshould produce a delayed homeostatic responsein visual cortex, whereas blocking optic nerveactivity or placing animals in the dark producesan immediate homeostatic response, remains amystery.

Currently we know little about the cellularand molecular mechanisms underlying homeo-static plasticity in vivo. Visual deprivation in-duced by optic nerve block scale ups mEPSCamplitude in vivo as in vitro, and results in in-sertion of GluA2-containing AMPAR (Desaiet al. 2002; Maffei and Turrigiano 2008; Gaineyet al. 2009). In contrast, dark rearing increasesthe abundance of GluA2-lacking receptors andrequires GluA1 phosphorylation, suggestingthat these two modes of deprivation may inducedifferent forms of compensatory synaptic plas-ticity (Goel and Lee 2007; Goel et al. 2011). In-terestingly, the delayed potentiation in visualcortex following lid suture requires TNFa andArc signaling (Kaneko et al. 2008; McCurryet al. 2011) as does synaptic scaling in vitro(Shepherd et al. 2006; Stellwagen and Malenka2006; Steinmetz and Turrigiano 2010); however,

both these manipulations likely affect manyforms of plasticity, so it still remains to beseen if synaptic scaling contributes to the de-layed potentiation induced by prolonged lidsuture. Taken together, the findings describedabove suggest that something very like synapticscaling operates in vivo in the intact visual cor-tex, but further studies are needed to verify thatwhat is observed in vivo is mechanistically iden-tical to in vitro synaptic scaling.

It should be remembered that homeostaticresponses in vivo often couple scaling of excita-tory synapses with the selective modificationof inhibitory networks, so that there is anoverall rebalancing of excitation and inhibition(Turrigiano and Nelson 2004). This highlightsthe idea that experience-dependent plasticityis unlikely to be explained by a single form ofsynaptic plasticity, but rather arises through acomplex interplay between many forms ofexcitatory, inhibitory, and also intrinsic plasti-city mechanisms occurring at many sites withinthe cortical microcircuit (Turrigiano 2011).Understanding experience-dependent plasticitywill thus require an integrated understanding ofhow these various forms of plasticity cooperateto modify microcircuit function. The existenceof this rich palette of plasticity mechanisms sug-gests that cortical microcircuits can respond in avery flexible manner to changes in sensory in-put. In particular, the existence of many formsof homeostatic plasticity operating on differenttemporal and spatial scales may ensure that net-work compensation can be achieved in responseto a wide range of sensory perturbations.

CONCLUDING REMARKS

Neurons and networks use a family of homeo-static synaptic plasticity mechanisms to stabilizefiring rates in the face of developmental orlearning-induced changes in drive, and thiscontributes to the ability of central neuronalnetworks to maintain stable function and en-ables networks to maintain the specificityof syn-aptic changes that encode information. Fromconsiderable work on the activity sensors, ex-pression mechanisms, and signaling pathwaysinvolved, the field is beginning to illuminate





how homeostatic negative feedback systems aredesigned and are able to control various aspectsof synaptic function. There appear to be inde-pendent mechanisms for regulating presynapticand postsynaptic strength, and there is growingevidence for independent homeostatic mech-anisms operating on global and local spatialscales. As additional phenomena are uncoveredthat appear to operate in a homeostatic or adap-tive manner, it will become increasingly impor-tant to begin to unravel and differentiate theroles these mechanisms play in tuning synaptictransmission, and how they influence informa-tion flow and storage in the intact central ner-vous system.

REFERENCES

Abbott LF, Nelson SB. 2000. Synaptic plasticity: Taming thebeast. Nat Neurosci 3: 1178–1183.

Anggono V, Clem RL, Huganir RL. 2011. PICK1 loss offunction occludes homeostatic synaptic scaling. J Neuro-sci 31: 2188–2196.

Aoto J, Nam CI, Poon MM, Ting P, Chen L. 2008. Synapticsignaling by all-trans retinoic acid in homeostatic synap-tic plasticity. Neuron 60: 308–320.

Bartley AF, Huang ZJ, Huber KM, Gibson JR. 2008. Dif-ferential activity-dependent, homeostatic plasticity oftwo neocortical inhibitory circuits. J Neurophysiol 100:1983–1994.

Bats C, Groc L, Choquet D. 2007. The interaction betweenStargazin and PSD-95 regulates AMPA receptor surfacetrafficking. Neuron 53: 719–734.

Beattie EC, Stellwagen D, Morishita W, Bresnahan JC, HaBK, Von Zastrow M, Beattie MS, Malenka RC. 2002.Control of synaptic strength by glial TNFa. Science 295:2282–2285.

Berg DK, Hall ZW. 1975. Increased extrajunctional acetyl-choline sensitivity produced by chronic acetylcholinesensitivity produced by chronic post-synaptic neuromus-cular blockade. J Physiol 244: 659–676.

Bessis A, Bechade C, Bernard D, Roumier A. 2007. Micro-glial control of neuronal death and synaptic properties.Glia 55: 233–238.

Bienenstock EL, Cooper LN, Munro PW. 1982. Theory forthe development of neuron selectivity: Orientationspecificity and binocular interaction in visual cortex.J Neurosci 2: 32–48.

Branco T, Staras K, Darcy KJ, Goda Y. 2008. Local dendriticactivity sets release probability at hippocampal synapses.Neuron 59: 475–485.

Burrone J, O’Byrne M, Murthy VN. 2002. Multiple formsof synaptic plasticity triggered by selective suppressionof activity in individual neurons. Nature 420: 414–418.

Cannon WB. 1932. The wisdom of the body. W.W. Norton,New York.

Chandrasekaran AR, Shah RD, Crair MC. 2007. Develop-mental homeostasis of mouse retinocollicular synapses.J Neurosci 27: 1746–1755.

Chang MC, Park JM, Pelkey KA, Grabenstatter HL, Xu D,Linden DJ, Sutula TP, McBain CJ, Worley PF. 2010.Narp regulates homeostatic scaling of excitatory synapseson parvalbumin-expressing interneurons. Nat Neurosci13: 1090–1097.

Chowdhury S, Shepherd JD, Okuno H, Lyford G, PetraliaRS, Plath N, Kuhl D, Huganir RL, Worley PF. 2006.Arc/Arg3.1 interacts with the endocytic machinery toregulate AMPA receptor trafficking. Neuron 52: 445–459.

Cingolani LA, Thalhammer A, Yu LM, Catalano M, RamosT, Colicos MA, Goda Y. 2008. Activity-dependent regula-tion of synaptic AMPA receptor composition and abun-dance by b3 integrins. Neuron 58: 749–762.

Copi A, Jungling K, Gottmann K. 2005. Activity- andBDNF-induced plasticity of miniature synaptic currentsin ES cell-derived neurons integrated in a neocortical net-work. J Neurophysiol 94: 4538–4543.

Davis GW. 2006. Homeostatic control of neural activity:From phenomenology to molecular design. Annu RevNeurosci 29: 307–323.

Davis GW, Bezprozvanny I. 2001. Maintaining the stabilityof neural function: A homeostatic hypothesis. AnnuRev Physiol 63: 847–869.

Davis GW, Goodman CS. 1998. Synapse-specific control ofsynaptic efficacy at the terminals of a single neuron.Nature 392: 82–86.

Deeg KE, Aizenman CD. 2011. Sensory modality-specifichomeostatic plasticity in the developing optic tectum.Nat Neurosci 14: 548–550.

De Gois S, Schafer MK, Defamie N, Chen C, Ricci A, WeiheE, Varoqui H, Erickson JD. 2005. Homeostatic scaling ofvesicular glutamate and GABA transporter expression inrat neocortical circuits. J Neurosci 25: 7121–7133.

Desai NS, Cudmore RH, Nelson SB, Turrigiano GG. 2002.Critical periods for experience-dependent synapticscaling in visual cortex. Nat Neurosci 5: 783–789.

Echegoyen J, Neu A, Graber KD, Soltesz I. 2007. Homeo-static plasticity studied using in vivo hippocampalactivity-blockade: Synaptic scaling, intrinsic plasticityand age-dependence. PLoS One 2: e700.

Ehlers MD, Heine M, Groc L, Lee MC, Choquet D. 2007.Diffusional trapping of GluR1 AMPA receptors by input-specific synaptic activity. Neuron 54: 447–460.

Ehrlich I, Malinow R. 2004. Postsynaptic density 95 controlsAMPA receptor incorporation during long-term poten-tiation and experience-driven synaptic plasticity. J Neuro-sci 24: 916–927.

Ehrlich I, Klein M, Rumpel S, Malinow R. 2007. PSD-95 isrequired for activity-driven synapse stabilization. ProcNatl Acad Sci 104: 4176–4181.

Erickson JD, De Gois S, Varoqui H, Schafer MK, Weihe E.2006. Activity-dependent regulation of vesicular gluta-mate and GABA transporters: A means to scale quantalsize. Neurochem Int 48: 643–649.

Evers DM, Matta JA, Hoe HS, Zarkowsky D, Lee SH, Isaac JT,Pak DT. 2010. Plk2 attachment to NSF induces homeo-static removal of GluA2 during chronic overexcitation.Nat Neurosci 13: 1199–1207.

G. Turrigiano




Frank CA, Kennedy MJ, Goold CP, Marek KW, Davis GW.2006. Mechanisms underlying the rapid induction andsustained expression of synaptic homeostasis. Neuron52: 663–677.

Frank CA, Pielage J, Davis GW. 2009. A presynaptic homeo-static signaling system composed of the Eph receptor,ephexin, Cdc42, and Cav2.1 calcium channels. Neuron61: 556–569.

Frenkel MY, Bear MF. 2004. How monocular deprivationshifts ocular dominance in visual cortex of young mice.Neuron 44: 917–923.

Fu AK, Hung KW, Fu WY, Shen C, Chen Y, Xia J, Lai KO,Ip NY. 2011. APC(Cdh1) mediates EphA4-dependentdownregulation of AMPA receptors in homeostaticplasticity. Nat Neurosci 14: 181–189.

Gainey MA, Hurvitz-Wolff JR, Lambo ME, Turrigiano GG.2009. Synaptic scaling requires the GluR2 subunit of theAMPA receptor. J Neurosci 29: 6479–6489.

Goddard CA, Butts DA, Shatz CJ. 2007. Regulation of CNSsynapses by neuronal MHC class I. Proc Natl Acad Sci104: 6828–6833.

Goel A, Lee HK. 2007. Persistence of experience-inducedhomeostatic synaptic plasticity through adulthood insuperficial layers of mouse visual cortex. J Neurosci 27:6692–6700.

Goel A, Xu LW, Snyder KP, Song L, Goenaga-Vazquez Y,Megill A, Takamiya K, Huganir RL, Lee HK. 2011. Phos-phorylation of AMPA receptors is required for sensorydeprivation-induced homeostatic synaptic plasticity.PLoS One 6: e18264.

Gonzalez-Islas C, Wenner P. 2006. Spontaneous network ac-tivity in the embryonic spinal cord regulates AMPAergicand GABAergic synaptic strength. Neuron 49: 563–575.

Goold CP, Davis GW. 2007. The BMP ligand Gbb gates theexpression of synaptic homeostasis independent of syn-aptic growth control. Neuron 56: 109–123.

Goold CP, Nicoll RA. 2010. Single-cell optogenetic excita-tion drives homeostatic synaptic depression. Neuron 68:512–528.

Hanson MG, Landmesser LT. 2004. Normal patterns ofspontaneous activity are required for correct motoraxon guidance and the expression of specific guidancemolecules. Neuron 43: 687–701.

Hardingham NR, Read JC, Trevelyan AJ, Nelson JC, Jack JJ,Bannister NJ. 2010. Quantal analysis reveals a functionalcorrelation between presynaptic and postsynaptic effi-cacy in excitatory connections from rat neocortex.J Neurosci 30: 1441–1451.

Harms KJ, Craig AM. 2005. Synapse composition and or-ganization following chronic activity blockade in cul-tured hippocampal neurons. J Comp Neurol 490: 72–84.

Hartman KN, Pal SK, Burrone J, Murthy VN. 2006. Activity-dependent regulation of inhibitory synaptic transmissionin hippocampal neurons. Nat Neurosci 9: 642–649.

Hou Q, Zhang D, Jarzylo L, Huganir RL, Man HY. 2008.Homeostatic regulation of AMPA receptor expression atsingle hippocampal synapses. Proc Natl Acad Sci 105:775–780.

Hu JH, Park JM, Park S, Xiao B, Dehoff MH, Kim S, HayashiT, Schwarz MK, Huganir RL, Seeburg PH, et al. 2010.

Homeostatic scaling requires group I mGluR activationmediated by Homer1a. Neuron 68: 1128–1142.

Huang ZJ, Kirkwood A, Pizzorusso T, Porciatti V, Morales B,Bear MF, Maffei L, Tonegawa S. 1999. BDNF regulates thematuration of inhibition and the critical period of plas-ticity in mouse visual cortex. Cell 98: 739–755.

Huupponen J, Molchanova SM, Taira T, Lauri SE. 2007. Sus-ceptibility for homeostatic plasticity is down-regulated inparallel with maturation of the rat hippocampal synapticcircuitry. J Physiol 581: 505–514.

Ibata K, Sun Q, Turrigiano GG. 2008. Rapid synaptic scalinginduced by changes in postsynaptic firing. Neuron 57:819–826.

Jakawich SK, Nasser HB, Strong MJ, McCartney AJ, PerezAS, Rakesh N, Carruthers CJ, Sutton MA. 2010. Localpresynaptic activity gates homeostatic changes in pre-synaptic function driven by dendritic BDNF synthesis.Neuron 68: 1143–1158.

Ju W, Morishita W, Tsui J, Gaietta G, Deerinck TJ, AdamsSR, Garner CC, Tsien RY, Ellisman MH, Malenka RC.2004. Activity-dependent regulation of dendritic synthe-sis and trafficking of AMPA receptors. Nat Neurosci 7:244–53.

Kaneko M, Stellwagen D, Malenka RC, Stryker MP. 2008.Tumor necrosis factor-a mediates one component ofcompetitive, experience-dependent plasticity in develop-ing visual cortex. Neuron 58: 673–680.

Kay L, Humphreys L, Eickholt BJ, Burrone J. 2011. Neuronalactivity drives matching of pre- and postsynaptic func-tion during synapse maturation. Nat Neurosci 14: 688–690.

Kilman V, van Rossum MC, Turrigiano GG. 2002. Activitydeprivation reduces miniature IPSC amplitude by de-creasing the number of postsynaptic GABA(A) receptorsclustered at neocortical synapses. J Neurosci 22: 1328–1337.

Kim J, Tsien RW. 2008. Synapse-specific adaptations to inac-tivity in hippocampal circuits achieve homeostatic gaincontrol while dampening network reverberation. Neuron58: 925–937.

Kirov SA, Sorra KE, Harris KM. 1999. Slices have moresynapses than perfusion-fixed hippocampus from bothyoung and mature rats. J Neurosci 19: 2876–2886.

Knogler LD, Liao M, Drapeau P. 2010. Synaptic scaling andthe development of a motor network. J Neurosci 30:8871–8881.

Leslie KR, Nelson SB, Turrigiano GG. 2001. Postsynaptic de-polarization scales quantal amplitude in cortical pyrami-dal neurons. J Neurosci 21: RC170.

Lu B. 2003. BDNF and activity-dependent synaptic modula-tion. Learn Mem 10: 86–98.

Maffei A, Nataraj K, Nelson SB, Turrigiano GG. 2006. Poten-tiation of cortical inhibition by visual deprivation.Nature 443: 81–84.

Maffei A, Turrigiano GG. 2008. Multiple modes of networkhomeostasis in visual cortical layer 2/3. J Neurosci 28:4377–4384.

Maffei A, Nelson SB, Turrigiano GG. 2004. Selective recon-figuration of layer 4 visual cortical circuitry by visual dep-rivation. Nat Neurosci 7: 1353–1359.





Malenka RC, Bear MF. 2004. LTP and LTD: An embarrass-ment of riches. Neuron 44: 5–21.

Marder E, Goaillard JM. 2006. Variability, compensationand homeostasis in neuron and network function. NatRev Neurosci 7: 563–574.

Marder E, Prinz AA. 2003. Current compensation in neu-ronal homeostasis. Neuron 37: 2–4.

McCurry CL, Shepherd JD, Tropea D, Wang KH, Bear MF,Sur M. 2011. Loss of Arc renders the visual cortex imper-vious to the effects of sensory experience or deprivation.Nat Neurosci 13: 450–457.

Miller KD, MacKay DJC. 1994. The role of constraints inHebbian learning. Neural Computation 6: 100–124.

Mrsic-Flogel TD, Hofer SB, Ohki K, Reid RC, Bonhoeffer T,Hubener M. 2007. Homeostatic regulation of eye-specificresponses in visual cortex during ocular dominance plas-ticity. Neuron 54: 961–972.

Nelson SB, Turrigiano GG. 2008. Strength through diversity.Neuron 60: 477–482.

O’Brien RJ, Kamboj S, Ehlers MD, Rosen KR, FischbachGD, Huganir RL. 1998. Activity-dependent modulationof synaptic AMPA receptor accumulation. Neuron 21:1067–1078.

Paradis S, Sweeney ST, Davis GW. 2001. Homeostatic controlof presynaptic release is triggered by postsynaptic mem-brane depolarization. Neuron 30: 737–749.

Perez-Otano I, Ehlers MD. 2005. Homeostatic plasticityand NMDA receptor trafficking. Trends Neurosci 28:229–238.

Poon MM, Chen L. 2008. Retinoic acid-gated sequence-specific translational control by RARa. Proc Natl AcadSci 105: 20303–20308.

Pozo K, Goda Y. 2010. Unraveling mechanisms of homeo-static synaptic plasticity. Neuron 66: 337–351.

Rabinowitch I, Segev I. 2006a. The endurance and selectiv-ity of spatial patterns of long-term potentiation/depres-sion in dendrites under homeostatic synaptic plasticity.J Neurosci 26: 13474–13484.

Rabinowitch I, Segev I. 2006b. The interplay between ho-meostatic synaptic plasticity and functional dendriticcompartments. J Neurophysiol 96: 276–283.

Rich MM, Wenner P. 2007. Sensing and expressing homeo-static synaptic plasticity. Trends Neurosci 30: 119–125.

Rutherford LC, DeWan A, Lauer HM, Turrigiano GG. 1997.Brain-derived neurotrophic factor mediates the activity-dependent regulation of inhibition in neocortical cul-tures. J Neurosci 17: 4527–4535.

Rutherford LC, Nelson SB, Turrigiano GG. 1998. BDNFhas opposite effects on the quantal amplitude of pyrami-dal neuron and interneuron excitatory synapses. Neuron21: 521–530.

Seeburg DP, Sheng M. 2008. Activity-induced Polo-like ki-nase 2 is required for homeostatic plasticity of hippocam-pal neurons during epileptiform activity. J Neurosci 28:6583–6591.

Seeburg DP, Feliu-Mojer M, Gaiottino J, Pak DT, Sheng M.2008. Critical role of CDK5 and Polo-like kinase 2 inhomeostatic synaptic plasticity during elevated activity.Neuron 58: 571–583.

Sharpless SK. 1975. Supersensitivity-like phenomena in thecentral nervous system. Fed Proc 34: 1990–1997.

Shepherd JD, Rumbaugh G, Wu J, Chowdhury S, Plath N,Kuhl D, Huganir RL, Worley PF. 2006. Arc/Arg3.1 medi-ates homeostatic synaptic scaling of AMPA receptors.Neuron 52: 475–484.

Simon DJ, Madison JM, Conery AL, Thompson-Peer KL,Soskis M, Ruvkun GB, Kaplan JM, Kim JK. 2008. The mi-croRNA miR-1 regulates a MEF-2-dependent retrogradesignal at neuromuscular junctions. Cell 133: 903–915.

Stein V, House DR, Bredt DS, Nicoll RA. 2003. Postsynapticdensity-95 mimics and occludes hippocampal long-termpotentiation and enhances long-term depression. J Neu-rosci 23: 5503–5506.

Steinberg JP, Takamiya K, Shen Y, Xia J, Rubio ME, Yu S, JinW, Thomas GM, Linden DJ, Huganir RL. 2006. Targetedin vivo mutations of the AMPA receptor subunit GluR2and its interacting protein PICK1 eliminate cerebellarlong-term depression. Neuron 49: 845–860.

Steinmetz CC, Turrigiano GG. 2010. Tumor necrosisfactor-a signaling maintains the ability of cortical synap-ses to express synaptic scaling. J Neurosci 30: 14685–14690.

Stellwagen D, Malenka RC. 2006. Synaptic scaling mediatedby glial TNF-a. Nature 440: 1054–1059.

Stellwagen D, Beattie EC, Seo JY, Malenka RC. 2005. Differ-ential regulation of AMPA receptor and GABA receptortrafficking by tumor necrosis factor-a. J Neurosci 25:3219–3228.

Sun Q, Turrigiano GG. 2011. PSD-95 and PSD-93 play crit-ical but distinct roles in synaptic scaling up and down.J Neurosci 31: 6800–6808.

Sun X, Wolf ME. 2009. Nucleus accumbens neurons exhibitsynaptic scaling that is occluded by repeated dopaminepre-exposure. Eur J Neurosci 30: 539–550.

Sutton MA, Ito HT, Cressy P, Kempf C, Woo JC, SchumanEM. 2006. Miniature neurotransmission stabilizes synap-tic function via tonic suppression of local dendritic pro-tein synthesis. Cell 125: 785–799.

Sutton MA, Taylor AM, Ito HT, Pham A, Schuman EM.2007. Postsynaptic decoding of neural activity: eEF2 asa biochemical sensor coupling miniature synaptic trans-mission to local protein synthesis. Neuron 55: 648–661.

Swanwick CC, Murthy NR, Kapur J. 2006. Activity-depend-ent scaling of GABAergic synapse strength is regulated bybrain-derived neurotrophic factor. Molec Cell Neurosci31: 481–492.

Terashima A, Pelkey KA, Rah JC, Suh YH, Roche KW, Col-lingridge GL, McBain CJ, Isaac JT. 2008. An essentialrole for PICK1 in NMDA receptor-dependent bidirec-tional synaptic plasticity. Neuron 57: 872–882.

Thiagarajan TC, Lindskog M, Tsien RW. 2005. Adaptation tosynaptic inactivity in hippocampal neurons. Neuron 47:725–737.

Thiagarajan TC, Lindskog M, Malgaroli A, Tsien RW. 2007.LTP and adaptation to inactivity: Overlapping mecha-nisms and implications for metaplasticity. Neuropharma-cology 52: 156–175.

Turrigiano GG. 2008. The self-tuning neuron: Synaptic scal-ing of excitatory synapses. Cell 135: 422–435.

G. Turrigiano




Turrigiano G. 2011. Too many cooks? Intrinsic and synaptichomeostatic mechanisms in cortical circuit refinement.Annu Rev Neurosci 34: 89–103.

Turrigiano GG, Nelson SB. 2004. Homeostatic plasticity in thedeveloping nervous system. Nat Rev Neurosci 5: 97–107.

Turrigiano GG, Leslie KR, Desai NS, Rutherford LC, NelsonSB. 1998. Activity-dependent scaling of quantal ampli-tude in neocortical neurons. Nature 391: 892–896.

Vale C, Sanes DH. 2002. The effect of bilateral deafness onexcitatory and inhibitory synaptic strength in the inferiorcolliculus. Eur J Neurosci 16: 2394–2404.

Volk L, Kim CH, Takamiya K, Yu Y, Huganir RL. 2010. De-velopmental regulation of protein interacting with Ckinase 1 (PICK1) function in hippocampal synaptic plas-ticity and learning. Proc Natl Acad Sci 107: 21784–21789.

Watt AJ, van Rossum MC, MacLeod KM, Nelson SB, Turri-giano GG. 2000. Activity coregulates quantal AMPA and

NMDA currents at neocortical synapses. Neuron 26:659–670.

Wierenga CJ, Ibata K, Turrigiano GG. 2005. Postsynaptic ex-pression of homeostatic plasticity at neocortical synapses.J Neurosci 25: 2895–2905.

Wierenga CJ, Walsh MF, Turrigiano GG. 2006. Temporal reg-ulation of the expression locus of homeostatic plasticity.J Neurophysiol 96: 2127–2133.

Xu W, Schluter OM, Steiner P, Czervionke BL, Sabatini B,Malenka RC. 2008. Molecular dissociation of the roleof PSD-95 in regulating synaptic strength and LTD. Neu-ron 57: 248–262.

Yu LM, Goda Y. 2009. Dendritic signalling and homeostaticadaptation. Curr Opin Neurobiol 19: 327–335.

Zhang W, Linden D.J.. 2003. The other side of the engram:Experience-driven changes in neuronal intrinsic excit-ability. Nat Rev Neurosci 4: 885–900.





November 15, 20112012; doi: 10.1101/cshperspect.a005736 originally published onlineCold Spring Harb Perspect Biol

Gina Turrigiano Stabilizing Neuronal FunctionHomeostatic Synaptic Plasticity: Local and Global Mechanisms for

Subject Collection The Synapse

SpinesStudying Signal Transduction in Single Dendritic

Ryohei Yasuda

Synaptic Vesicle EndocytosisYasunori Saheki and Pietro De Camilli

Synaptic Vesicle Pools and DynamicsAbdulRasheed A. Alabi and Richard W. Tsien

Short-Term Presynaptic PlasticityWade G. Regehr

Synapses and Memory Storage

KandelMark Mayford, Steven A. Siegelbaum and Eric R.

(LTP/LTD)Potentiation and Long-Term Depression NMDA Receptor-Dependent Long-Term

Christian Lüscher and Robert C. MalenkaSynapses and Alzheimer's Disease

C. SüdhofMorgan Sheng, Bernardo L. Sabatini and Thomas Brain

Ultrastructure of Synapses in the Mammalian

Kristen M. Harris and Richard J. WeinbergSynaptic Cell Adhesion

BiedererMarkus Missler, Thomas C. Südhof and Thomas

Calcium Signaling in Dendritic SpinesMichael J. Higley and Bernardo L. Sabatini

DisabilitiesIntellectualDisorders Associated with Autism and

Synaptic Dysfunction in Neurodevelopmental

Huda Y. Zoghbi and Mark F. Bear

Synaptic Neurotransmitter-Gated ReceptorsTrevor G. Smart and Pierre Paoletti

The Postsynaptic Organization of SynapsesMorgan Sheng and Eunjoon Kim

Synaptic Vesicle ExocytosisThomas C. Südhof and Josep Rizo

Inhibitory SynapsesPresynaptic LTP and LTD of Excitatory and

Pablo E. CastilloNeurotransmittersVesicular and Plasma Membrane Transporters for

Randy D. Blakely and Robert H. Edwards

http://cshperspectives.cshlp.org/cgi/collection/ For additional articles in this collection, see

Copyright © 2012 Cold Spring Harbor Laboratory Press; all rights reserved


http://cshperspectives.cshlp.org/cgi/collection/

http://cshperspectives.cshlp.org/cgi/collection/


Homeostatic Synaptic Plasticity: Local and Global ... · Homeostasis of neuronal ﬁring through...

Documents

Transcript of Homeostatic Synaptic Plasticity: Local and Global ... · Homeostasis of neuronal ﬁring through...