In higher vertebrates, the extracellular matrix (ECM) wrapping cells of the adult brain differs significantly from that of the developing and juvenile brain. The mature ECM is established at the end of critical periods for wiring and it restricts the regenerative potential and constrains the plasticity of the adult brain. In particular, perineuronal nets, elaborate ECM structures that surround distinct neurons and wrap synapses, are hallmarks of the adult brain and seem to massively affect brain plasticity. Why have these, at first glance futile, limitations evolved? What is the return for these drawbacks? What are the mechanisms of restriction and how is adult plasticity implemented? Recent progress both at the systemic level and at the molecular physiological level has shed some new light on these questions. In this review we will survey the evidence for potential functions of the adult ECM in making established brain features, including imprinted memories, resistant to extinction, and we will discuss potential mechanisms by which the ECM limits juvenile and implements adult plasticity. In particular we will focus on some aspects of adult ECM function. First we will discuss its influence on diffusion of cations in the extracellular space and on volume transmission, second we will consider its potential role in regulating the lateral diffusion of cell surface receptors and finally we will discuss mechanisms to locally modulate ECM functions.
The space between neural cells in the brain is filled with material of the extracellular matrix (ECM). Both neurons and glial cells contribute to the production of ECM components and the ECM in turn mediates various structural and functional interactions between these cells (Faissner et al., 2010). A basic component of the brain’s ECM is the unbranched polysaccharide hyaluronic acid that acts as backbone to noncovalently recruit proteoglycans and glycoproteins into ECM structures (Bandtlow & Zimmermann, 2000; Rauch, 2004; Frischknecht & Seidenbecher, 2008). Major constituents of the hyaluronan-based ECM are chondroitin sulfate proteoglycans (CSPGs) of the lectican/hyalectan family, tenascins and so-called link proteins (Bandtlow & Zimmermann, 2000; Yamaguchi, 2000; Rauch, 2004). In addition, a large variety of other components including reelin, laminins, pentraxins, pleiotrophin/HB-GAM, phosphocan, thrombospondins and heparan-sulfate proteoglycans (HSPGs), such as agrin or cell surface-bound HSPGs of the syndecan and glypican family, as well as the matrix-shaping enzymes such as proteases and hyaluronidases, contribute to the brain’s ECM (Bandtlow & Zimmermann, 2000; Dityatev & Schachner, 2003; Christopherson et al., 2005; Dityatev & Fellin, 2008; Frischknecht & Seidenbecher, 2008). The ECM undergoes significant changes during CNS development. In the mammalian brain, initially a juvenile form of the ECM is synthesized during late embryonic and early postnatal development. For example, the lecticans neurocan and distinct isoforms of versican (V0 and V1) as well as tenascin-C are prominent constituents of this juvenile matrix, whereas the adult ECM is characterized by down-regulation of these components and the up-regulation of other markers including the lecticans brevican, versican V2 and aggrecan as well as phosphocan and tenascin-R (Milev et al., 1998; Carulli et al., 2007; Zimmermann & Dours-Zimmermann, 2008).
The ECM of the embryonic and juvenile brain is permissive and supportive for neurogenesis and gliogenesis, cell migration, axonal outgrowth and axonal pathfinding, as well as for synaptogenesis and synaptic rearrangements (Bandtlow & Zimmermann, 2000; Faissner et al., 2010). In contrast, the adult ECM is nonpermissive for axonal outgrowth and inhibits regeneration and major reorganization processes in the adult CNS (Galtrey & Fawcett, 2007; Fawcett, 2009). In addition to a variety of other factors, CSPGs of the lectican family including brevican display an inhibitory activity on neurite outgrowth (Zurn & Bandtlow, 2006; Quaglia et al., 2008), and the removal of ECM by chondriotinase ABC constitutes a way to promote functional recovery in the injured brain (Crespo et al., 2007; Galtrey & Fawcett, 2007). The implementation of the adult ECM coincides with the end of the critical period (Lander et al., 1997; Fawcett, 2009), the time window during which neuronal circuits are shaped and refined by experience. The critical periods of enhanced structural and functional synaptic plasticity differ from brain area to brain area and from species to species, and can last for months to years in primates including humans (Hensch, 2004).
The restriction in regenerative and reorganizational plasticity of the CNS, which has evolved in higher vertebrates only, must provide an evolutionary advantage over lower vertebrates, which have largely retained this plasticity. This evolutionary benefit may include the suppression of regenerative processes that are time- and energy-consumptive and though beneficial for the individual not helpful for the survival of the population, and/or the preservation of costly acquired hardwired connections that are essential for the rapid experience-based processing of information in complex nervous systems. Nonetheless the adult nervous system retains a remarkable synaptic plasticity that is partly based on the local restoration of a ‘juvenile’ environment. Here, we will briefly survey the present knowledge about structure and functions of the adult ECM and then discuss potential mechanisms by which the adult ECM restricts juvenile synaptic plasticity and how this plasticity may be locally restored by the release or activation of ECM-removing or -modifying enzymes.
Structural features of the ECM in the brain
The brain’s ECM has a complex history. Although perineuronal nets (PNNs) were discovered, as prominent structures surrounding neurons, by the pioneers of brain cell biology including Camillo Golgi and Santiago Ramon y Cajal (for review see Celio et al., 1998), the existence of an ECM in the brain was only acknowledged in the 1970s (for review see Zimmermann & Dours-Zimmermann, 2008). Utilizing biochemical fractionation techniques it has been recognized that ECM components such as brevican tightly associate with synaptic protein preparations (Seidenbecher et al., 1995, 2002; Li et al., 2004). A systematic analysis of the rat ECM revealed various extractable fractions from the adult brain (Deepa et al., 2006). While most of the material is loosely associated with brain membranes, another fraction can be extracted by treatment with nonionic detergent and salt and is thought to be associated with neural cell membranes. A final fraction comprising roughly a quarter of the CSPG material and including brevican, neurocan, versican V2, aggrecan and phosphacan can only be extracted with urea. This fraction is not present in the young brain before closure of the critical period and is thought to represent cartilage-like ECM material forming the PNNs (Fawcett, 2009). This material can be entirely removed from brain structures using the hyaluronan hydrolyzing enzyme hyaluronidase and partly with chondroitinase ABC, an enzyme that removes glycosaminoglycan chains from CSPGs but can also display some hyaluronidase activity (Deepa et al., 2006).
Another interesting ECM component that may act directly at synapses is reelin, a large (∼ 400 kDa) glycoprotein that plays an important role in brain development, as competently reviewed on several occasions (e.g. Tissir & Goffinet, 2003; Forster et al., 2006). However, reelin is also involved in the maintenance of cortical lamination in the adult nervous system (Frotscher et al., 2009) and in processes of adult synaptic plasticity and pathology (Herz & Chen, 2006). The exact functional relation of reelin to classical PNN is currently unknown. However, in the adult forebrain reelin is expressed primarily by parvalbumin-negative interneurons that are not wrapped by prominent PNNs (Pesold et al., 1998, 1999). In addition some projection neurons in the cerebral cortex and excitatory granule cells in the cerebellum as well as distinct populations of neurons throughout the brain express reelin in the matured brain (Pesold et al., 1998; Ramos-Moreno et al., 2006).
Functions for the adult brain ECM
Various functions have been assigned to or proposed for the adult ECM (Table 1). These include the restriction of regenerative plasticity of the central nervous system but also the establishment of neuroprotective functions (Galtrey & Fawcett, 2007; Fawcett, 2009). Furthermore, components of the adult ECM such as brevican seem to be involved in tumor growth and tumor suppression (Gary et al., 1998; Sim et al., 2009). As ECM derivatives such as PNN and PNN-like structures are assembled from components synthesized by astrocytes and by neurons, they may serve important functions in neuron–glia interaction and communication. For example, ECM components play essential roles in the formation of myelin specializations (Susuki & Rasband, 2008). This interaction is primarily mediated via neurofascin-186. Also, via other cell surface receptors including CD44, the neural cell adhesion molecule NCAM and integrins, the ECM contacts cell surfaces, makes contact with specializations of the cortical cytoskeleton and thereby may serve mechanical stability and mediate or modulate signaling processes (Celio & Blumcke, 1994; Fox & Caterson, 2002; Dityatev & Schachner, 2003; Rauch, 2004; Frischknecht & Seidenbecher, 2008). ECM structures have been further discussed as low-affinity receptors for trophic and growth factors (Celio & Blumcke, 1994; Galtrey & Fawcett, 2007) and as regulators of extracellular ion homeostasis (Hartig et al., 1999; Hrabetova et al., 2009; see below).
Table 1. Functions of the adult extracellular matrix of the CNS
Restriction of CNS regenerative plasticity
Generation of nonpermissive substrate for neurite outgrowth; inhibition of axonal regeneration after nerve crush (CSPG/hyaluranon-based ECM) Inhibition of synapse formation Removal of adult ECM structures with chondroitinase ABC partly restores regenerative potential
By interacting with in channels and cell adhesion molecules and in turn cytoskeletal specializations ECM components serve important roles in the localization of ion channels at nodes of Ranvier and axon initial segments
A most fascinating aspect of adult ECM function might be to terminate the critical period of circuit wiring and to implement adult plasticity modes. As mentioned above, the appearance of PNN coincides with the termination of critical periods of experience-dependent brain wiring. Dark-rearing prolongs the critical period and postpones PNN formation in the visual cortex (Lander et al., 1997; Pizzorusso et al., 2002). Similarly, deprivation of excitatory neuronal activity seems to delay the development of PNNs (Reimers et al., 2007). For the visual cortex of rats the critical period ends ∼3 weeks after birth (Hensch, 2004). Experiments by Pizzorusso et al. (2002) have demonstrated that removal of the PNN-like ECM from the visual cortex can restore this type of plasticity. The authors injected chondroitinase ABC into the visual cortex and showed that monocular deprivation resulted in a shift in ocular dominance in a manner similar to that described in the classical Wiesel & Hubel (1965) experiments. Utilizing this treatment it was even possible to recover the normal ocular dominance and to restore visual acuity to adult animals which had grown up with one long-term deprived eye (Pizzorusso et al., 2006). These experiments strongly suggest that one important function of the adult ECM is to terminate juvenile plasticity and to fix acquired experience-dependent wiring for the adult life.
A more recent study by Gogolla et al. (2009) suggests that similar mechanisms may make particular memories, such as fear memories, erasure-resistant, i.e., insensitive to extinction. In young rats, conditioned fear memories can be erased permanently whereas rats older than 3–4 weeks are resistant to this fear extinction. Fear extinction in both adult and young rats is amygdala-dependent. In this brain structure, PNNs develop between postnatal days 16 and 21. After this critical period fear memory can be reduced by repeated exposure to the conditioned stimulus in the absence of the aversive fear-provoking stimulus. However, in contrast to young animals, fear response is reinstated when the aversive stimulus is presented again. Similar to the experiments in the visual cortex, removal of the hyaluronan–CSPG-based ECM achieved a rapid and permanent erasure of newly acquired fear memories. Extinction did not take place when fear experience took place before the application of chondroitinase, suggesting that CSPGs are essential for protecting fear memories from erasure during the acquisition phase (Gogolla et al., 2009; Pizzorusso, 2009).
The mechanisms by which the hyaluronan–CSPG-based ECM performs its functions in establishing adult CNS plasticity are still largely unknown. However, a number of studies suggest that the adult ECM is importantly involved in various aspects of synaptic plasticity, which may contribute to the observed phenomena. These aspects include mechanisms of classical (Hebbian) plasticity as well as homeostatic synaptic plasticity and metaplasticity (see Dityatev & Schachner, 2003; Dityatev & Fellin, 2008 for a comprehensive review). In essence, most functions of the ECM have been reviewed on numerous occasions (for an overview see Table 1). Therefore, for the purpose of this article we will focus in the following sections on few aspects of adult ECM functions that may be important for the understanding of the implementation of adult plasticity mechanisms in the CNS. These are the control of extracellular diffusion events and the control of lateral diffusion of plasma membrane proteins. Finally, we will consider mechanisms to locally modulate ECM functions.
Role of the ECM in the control of diffusible signals
The interneuronal communication within neuronal networks is dominated by the diffusive transmission of signaling molecules. In addition to electrical synapses formed by gap junctions or signaling through cell surface-anchored receptor-ligand systems, many other modes of cellular communications have to bridge the extracellular space (ECS) between adjacent cells utilizing diffusible factors. The chemical synapse is the most direct form of cellular communication between neurons; here, the exact apposition of pre- and postsynaptic membranes optimizes the success of intercellular communication via transmitter diffusion. Many other forms of cellular communication in the brain seem to rely on the diffusion properties of the ECS and the much less accurately defined positioning of signaling molecules in the neural cell membrane. This type of diffusible transmission is designated volume or extrasynaptic transmission. As described for calcium ions (Hrabetova et al., 2009), neurotransmitters (Scimemi & Beato, 2009) and proteins (Thorne et al., 2008) the diffusion properties depend on a variety of factors including temperature, viscosity, charge and shape of the ECS, collectively and formally characterized by the tortuosity (reviewed by Sykova & Nicholson, 2008). The ECM primarily determines the charge and viscosity of the ECS, whereas membrane protuberances of neurons and glial cells, such as spines and filopodia, cause the structural restrictions for free diffusion in the ECS, also defined as geometric tortuosity (Kullmann et al., 1999).
Regulation of extracellular ion homeostasis
Measurements of extracellular ion concentrations during neuronal activity have revealed changes in the relation between potassium, sodium, calcium and chloride ions during synaptic transmission (Heinemann et al., 1977; Rausche et al., 1990) that influence the membrane potential of the active cell population. Hence local ion fluxes can function as feedback mechanisms for the active population of synapses (Rusakov & Fine, 2003). The high content of negatively charged CSPGs in the ECM is very likely to affect local changes of ion concentrations. A recent study on diffusion properties of cations in the ECS suggests that negatively charged CSPGs change these diffusion properties in particular for calcium ions. By removing the charged chondroitin sulfate side chains with chondrotinase ABC, Hrabetova et al. (2009) were able to detect a global increase in the effective diffusion coefficient of bivalent ions such as calcium, whereas the diffusion properties of the monovalent cation tetraethylammonium did not change. Physiologically, a local depletion of extracellular calcium can occur as a result of the frequent activation of postsynaptic NMDA receptors and hence decrease the presynaptic release probability, as demonstrated for the CA3 mossy fiber synapse in the hippocampus (Rusakov & Fine, 2003).
The ECM density is particularly high in the PNN around GABAergic, parvalbumin-containing fast-spiking interneurons. Because of their high negative charge, Hartig et al. (1999, 2001) postulated that one function of the PNN might be to increase the local ion buffer capacity in order to balance local depletion of cations during high-frequency firing activity. The observation of unaltered diffusion properties of monovalent cations after removal of the ECM with chondroitinase ABC (Hrabetova et al., 2009) does not, however, support this view of the ECM as a structure directly involved in the spatial buffering of monovalant cations. Recent progress in the understanding of neuron–glia and glia–vasculature communication rather highlights the special molecular properties of glial networks (Volterra & Meldolesi, 2005; Rouach et al., 2008; Giaume et al., 2010) and emphasizes a dominant role for neuron–glial interactions in the control of extracellular cation concentrations (Kofuji & Newman, 2004; Frohlich et al., 2008).
Another aspect of the ECM function as a structure modulating the excitability of the membrane is the involvement in the localization and membrane organization of voltage-gated ion channels as postulated by Kaplan et al. (1997). Tenascins R and C have been reported to interact directly with voltage-gated sodium channels. This interaction with the auxiliary β1 and β2 subunits modulates their subcellular localization during myelinization of the axonal membrane (Srinivasan et al., 1998; Xiao et al., 1999; Isom, 2001). Other ECM molecules including brevican may also contribute to the function of the ECM to induce and stabilize surface compartmentalization of signaling molecules and to organize and cluster ion-conducting protein complexes in the membrane of nodes of Ranvier (Susuki & Rasband, 2008). Further interactions between ECM components and ion channels were studied with respect to changes in gating and kinetic properties of potassium channels by the ECM component vitronectin (Vasilyev & Barish, 2003, 2004). Moreover, the modulation of L-type calcium channels by tenascins has profound influences on classical plasticity models, including long-term potentiation, long-term depression and metaplasticity (Evers et al., 2002; Dityatev & Schachner, 2003; Dityatev & Fellin, 2008). Hence, the ECM not only acts as a charged passive structure between neural cells but also actively modulates membrane conductance and excitability and contributes to the surface organization of signaling molecules including ion channels.
Modulation of neuron–glia interaction via ambient transmitter
Another important neuron-glia interaction is the modulation of neurotransmitter release and uptake, which modulates the activation of ionotropic and metabotropic receptors in both cell types inside and outside synapses. The time course of synaptic currents as well as the excitability of the postsynaptic neuron change during synaptogenesis for inhibitory and excitatory synapses in the CNS and in the peripheral nervous system. Various examples have been reported for developmental changes in presynaptic (Wasling et al., 2004) and postsynaptic molecular properties (Hestrin, 1992; Takahashi et al., 1992; Tia et al., 1996). Some synapses do not undergo major changes in their molecular assembly but experience drastic structural changes. An example is the mossy fiber–granular cell synapse of the cerebellum where AMPA receptor-mediated currents become accelerated during development due to a more focused transmitter release (Cathala et al., 2005). Gliagenesis and structural changes of the synapse itself and the surrounding neuropil lead to a faster rise and decay of miniature excitatory postsynaptic currents without changes in amplitude. The adult ECM as a negatively charged glue between astrocytes and neurons develops during the same time window (Bruckner et al., 2000; Carulli et al., 2006, 2007; Ishii & Maeda, 2008) and may further restrict glutamate diffusion. Sensing the distribution of activated AMPA receptors utilizing the low-affinity antagonist kynurenic acid confirmed the more focalized activation of receptors at the postsynaptic side in mature synapses (Cathala et al., 2005).
An impact of the local charge distribution on the diffusion properties of glutamate has recently been demonstrated by comparing AMPA receptor current decay time constants at negative and positive membrane potential (Sylantyev et al., 2008). The authors argue that transient events of depolarization during synaptic activity cause a positive net charge within the synaptic cleft that will prolong the dwell time of the negatively charged glutamate in this compartment. GABA as an electrically neutral transmitter does not display such effects (Sylantyev et al., 2008). Thus the electro-diffusion of glutamate modulates the AMPA receptor occupation as can be observed in the decay characteristics of the current. As a negatively charged structure, the hyaluronan–CSPG-based ECM, which does not penetrate the synaptic cleft, could accelerate the dispersion of glutamate once it leaves the cleft or it could contribute to the prolongation of the dwell-time of glutamate within the synaptic cleft by hindering diffusion of glutamate out of the cleft. Whether and how the ECM may influence the local concentration of ambient extrasynaptic glutamate is currently unknown. Another important parameter that we need to know to fully appreciate the complex scenario, the average concentration of ambient glutamate, is still a matter of debate (Bouvier et al., 1992; Herman & Jahr, 2007; Featherstone & Shippy, 2008).
Guidance of lateral diffusion of cell surface molecules by ECM-derived surface compartments
A large pool of surface molecules is highly mobile due to lateral Brownian diffusion within the plasma membrane (Kusumi et al., 1993; Triller & Choquet, 2008). In most cases, lateral diffusion of surface molecules is restricted by obstacles (pickets and corrals) compartmentalizing the cell surface, which may be formed by the underlying cytoskeleton or by rigid membrane structures (Kusumi et al., 1993, 2005). Although synapses only occupy a few per cent of the neuronal membrane surface, it is a subcellular compartment with an exceedingly important function in neurons as it is the main location for interneuronal neurotransmission. Neurotransmitter receptors, such as AMPA-type and NMDA-type glutamate receptors or GABAA receptors, are present not only in synaptic areas but also in extrasynaptic domains and lateral diffusion properties of receptors between these two domains have been investigated intensely. In general, diffusion of these receptors is more confined in the synaptic compartment than in extrasynaptic areas. However, receptors are steadily exchanging between the synaptic and extrasynaptic pools. This mechanism is probably fundamental for the maintenance of the synaptic receptor pool as the exchange between cell surface and intracellular receptors through exo- and endocytosis occurs outside the synaptic membrane (Newpher & Ehlers, 2008; Petrini et al., 2009). In addition, studies on hippocampal slices and primary hippocampal neurons have revealed that this lateral diffusion may account for the exchange of desensitized synaptic AMPA receptors, which emerge during high frequency firing, for naïve extrasynaptic ones (Heine et al., 2008). Blockade of lateral diffusion, e.g. by crosslinking with antibodies, resulted in strong paired-pulse depression (PPD) caused by desensitized receptors accumulating under the release site. These results demonstrated that the lateral diffusion of AMPA receptors was a novel postsynaptic mechanism influencing short-term plasticity of individual synapses.
Interestingly, the diffusion rates of AMPA receptors on dissociated hippocampal neurons decreased during synapse maturation, between the second and third week in vitro (Borgdorff & Choquet, 2002). During this time period, a hyaluronan–CSPG-based ECM resembling the perisynaptic net-like ECM of the adult CNS is formed in these cultures (John et al., 2006). Similar to the in vivo situation, the net-like structure divides the neuronal surface into multiple compartments of variable size (Fig. 1, see above). These ECM-derived cell surface structures restrict the lateral diffusion of extrasynaptic AMPA receptors (Frischknecht et al., 2009). Removal of the ECM with the enzyme hyaluronidase increased diffusion rates of extrasynaptic receptors and the exchange rate between synaptic and extrasynaptic receptors. This resembles the ‘juvenile’ situation before the ECM is established in the cultures (day 10 in vitro). An electrophysiological examination revealed that removal of ECM from dissociated hippocampal neurons affected short-term synaptic plasticity, i.e., in the presence of the ECM, PPD seems to be much stronger than after hyaluronidase treatment, when basically no PPD was observed. A similar down-regulation of AMPAR movement during synaptic maturation was observed when studying the role of stargazin in controlling AMPAR immobilization. Interestingly, overexpression in mature neurons of mutant stargazin unable to bind their intracellular partner PSD-95 reverted to the behavior of AMPAR to ‘juvenile’ type (Bats et al., 2007). Consequently, both intracellular and ECM-derived surface compartments can influence short-term plasticity of neurons by controlling lateral diffusion and thus control the synaptic availability of naïve AMPA receptors. It should be noted here that ECM nets are not impermeable barriers for diffusing surface proteins. They rather have to be considered as viscous structures that reduce the surface mobility of proteins. Accordingly, the size and shape of the extracellular domains of surface-exposed membrane proteins influences the mobility shift by the ECM (Frischknecht et al., 2009). Along this line, the recent characterization of the full crystal structure of AMPARs points to their very large extracellular domain, protruding over 10 nm into the extracellular space (Sobolevsky et al., 2009) and thus likely to bump into the ECM components.
As not only AMPA receptors but also the diffusion of other surface molecules with no known interaction partners, such as the glycosylphosphatidylinositol-anchored green-fluorescent protein, was affected by hyaluronidase treatment, we conclude that the hyaluronic acid-based ECM may function primarily as a passive diffusion barrier for cell surface proteins. However, in addition specific direct interactions of neurotransmitter receptors with components of the ECM may influence the lateral diffusion of neurotransmitter receptors in particular and cell surface receptors in general. One example of such interaction is the clustering of AMPA receptors by the pentraxin family member Narp (O’Brien et al., 1999). This immediate early gene forms clusters on the neuronal surface and co-aggregates AMPA receptors on spinal cord neurons.
The size of surface compartments and the density of the ECM meshwork may play an important role in controlling the access of AMPA receptors to the synapse und thus influence their synaptic properties. For example, ECM structures could regulate the accessibility of exocytic and endocytic sites for a distinct receptor population. Both size of compartments and density of the ECM may be actively regulated by neurons and astrocytes contributing to the synthesis of ECM material. On the one hand this may be achieved by altered expression of components of the ECM. In particular, the synthesis of hyaluronic acid seems to be crucial for the formation of the dense ECM (Carulli et al., 2006). Moreover, it has been reported that formation of PNN-like structures in vitro is regulated by activity and that the removal of the ECM altered interneuron activity (Dityatev et al., 2007). However, these effects are rather slow and may account for long-term changes in neuronal properties rather than fast changes.
On the other hand, local removal of ECM structures can contribute to plasticity regulation. Degradation of pre-existing ECM may be achieved on a much shorter time scale and could regulate both compartment size and the density of the ECM. A number of proteases, especially from the family of the matrix metalloproteases (MMPs) have been reported to cleave components of the ECM (Nakamura et al., 2000; Ethell & Ethell, 2007). In particular, ADAMTS4 (a disintegrin and metalloproteinase with thrombospondin motifs 4) has the potential to modify ECM structure as it degrades two of the major components of the hyaluronan–CSPG-based ECM, i.e. aggrecan and brevican. However, its impact on neuronal function and its regulation during neuronal activity remain to be clarified.
One of the best-studied MMPs in the nervous system is MMP9. Its activity has been associated with enhanced neuronal activity (Michaluk et al., 2007) and depletion of MMP9 results in impairment of long-term potentiation at hippocampal synapses (Nagy et al., 2006). Application of MMP9 to neuronal primary cultures affects lateral diffusion of NMDA receptors without changing the mobility of AMPA receptors or the structure of the hyaluronic acid-based ECM (Michaluk et al., 2009). Rather, extracellular MMP9 proteolysis induced integrin beta1-dependent signaling, which then led to the mobilization of NMDA receptors. Integrin beta1 signaling was also identified as being responsible for MMP9-induced spine enlargement and synaptic potentiation (Wang et al., 2008). To date the target of MMP9 proteolysis within the ECM that induces integrin signaling remains unknown.
Another matrix metalloprotease, MMP3, is able to process agrin at the basal lamina of the neuromuscular junction in an activity-dependent manner (Werle & VanSaun, 2003). It has been suggested that this process is responsible for fast removal of agrin from the neuromuscular junction, where it induces acetycholine receptor clustering and is indispensible for normal neuromuscular junction development (Sanes & Lichtman, 2001). In line with this view, MMP3-knockout mice exhibit abnormal neuromuscular junction morphology and acetylcholine receptor distribution (VanSaun et al., 2003). Another protease that has recently been found to act on agrin as its main substrate is the brain-specific serine protease neurotrypsin (Reif et al., 2008). Neurotrypsin can be released at synapses in an activity-dependent manner where it locally processes agrin into distinct fragments (Frischknecht et al., 2008; Stephan et al., 2008). Interestingly, neurotrypsin has been identified as essential for cognitive function in the human brain (Molinari et al., 2002). In an elegant series of experiments it was demonstrated that the proteolytic fragment of 22 kDa acquired signaling properties that induced filopodia in hippocampal slice cultures after induction of synaptic long-term potentiation (Matsumoto-Miyai et al., 2009). Thus, similar to MMP9, proteolytic cleavage of ECM components unmasked a signaling molecule, which in turn led to an altered spine morphology and even to the generation of new synapses. Hence, these examples demonstrate that the ECM contains a variety of hidden instructive signals that can be revealed by specific proteolysis.
Finally, it has been demonstrated that, similar to chondroitinase ABC treatment, the topical application of the serine protease tissue-type plasminogen activator (tPA) can prolong the so-called critical period in the visual cortex (Mataga et al., 2004; Oray et al., 2004). It was shown that tPA induces activity-dependent pruning of dendritic spines in the visual cortex. However, it is currently unclear whether pruning after tPA treatment depends on a newly generated signaling molecule, similar to the mode of action of neurotrypsin or MMP9, or whether the effect is based on the removal of the PNN-like structures as a general barrier for filopodial outgrowth (Berardi et al., 2004).
The ECM protein reelin has also been discussed as a serine protease (Quattrocchi et al., 2002). In the adult CNS it mediates its function via binding to its cell surface receptors very-low-density lipoprotein receptor (VLDLR) and ApoE2 receptor (APOE2R) and the downstream adaptor protein Dab1. These receptors seem to be directly in contact with synaptic NMDA receptors via the membrane-associated guanylate kinase homologue PSD-95, and reelin signaling is closely connected to NMDA receptor signaling and thereby regulates synaptic plasticity (for review see Herz & Chen, 2006; Rogers & Weeber, 2008). Reelin seems to exert important functions during the transition from the developing to the mature brain. Thus it has been implicated in the control of the subunit composition of somatic NMDA receptors during hippocampal maturation (Sinagra et al., 2005). Moreover, the same group reported recently that reelin secreted by GABAergic interneurons is responsible for maintaining the adult NMDA receptor composition and that blocking reelin secretion reversibly increases the fraction of juvenile NR2B-containing NMDA receptors. This effect can be rescued by supplementing exogenous reelin (Campo et al., 2009). Finally, reelin controls the surface trafficking of NR2B-containing NMDA receptors. As shown by single-particle tracking, inhibition of reelin function reduced the surface mobility of these receptors and increased their synaptic dwell time (Groc et al., 2007). This effect depended on beta1-containing integrin receptors, which are supposed to co-operate with APOE2Rs and/or VLDLRs. Currently it is unclear whether the protease activity of reelin plays a role in these processes.
Conclusions and perspectives
The ECM of the adult brain has features that differ considerably from those of the developing and the juvenile brain. Its implementation has dramatic consequences for the brain physiology. This becomes most obvious in the severe reduction of the regenerative potential that has long been recognized. Another feature to which the adult ECM contributes is the closure of the critical period, which may serve the stabilization of brain wiring after a period of experience-driven refinement. This has been impressively documented by the experiments of Pizzorusso et al. (2002) for the visual cortex. Recent experiments on the extinction of fear memories (Gogolla et al., 2009) suggest there is much more to be disclosed. These experiments suggest that memory acquisition differs between juvenile and adult brains and that adult structures of the hyaluronan–CSPG-based ECM are essential for an imprinted memory to bad experience. One does not have to be an augur to predict that we will face a multitude of studies that will unravel the function of PNNs and perisynaptic ECM structures in long-term memory processes.
As we have tried to illustrate in our article, the first details are emerging about how molecular and cellular mechanisms govern the adult ECM implementation of its functionality. A major principle seems to be to restrict lateral diffusion of cell surface molecules and to change the diffusion conditions, i.e. the tortuosity, for ions, small molecules and even macromolecules in the extracellular space. This in turn affects a large variety of parameters including calcium homeostasis, volume transmission of glutamate and other charged messengers, and local concentrations of signaling molecules. In particular the PNN-like perisynaptic ECM which, similar to the neuromuscular junction, is different from that in the synaptic cleft, seems to control the access to the synapse and to affect synaptic plasticity at various levels, from short-term plastic processes to homeostasis and metaplasticity. Many plastic processes employ ectoenzymes that may restore locally ‘juvenile’ environments in addition to generating new signaling molecules from cell surface and ECM products. The window for this type of research has just been opened and new views on basically important and medically relevant mechanisms of brain plasticity will emerge. These might include a deeper understanding of mental disorders including anxiety disorders (Pizzorusso, 2009), as well as schizophrenia and affective disorders that generally develop after the closure of major critical periods for higher brain functions of the prefrontal cortex after adolescence.
We wish to thank Dr Amin Derouiche, Bonn, for providing a photomicrograph for Fig. 1. Research in the authors’ laboratories on this topic is funded by the DFG (GU230/5-1,2,3; HE3604/2-1) and by ERA-Net NEURON (Moddifsyn).