The ocular dominance (OD) shift induced by monocular deprivation (MD) during the critical period is mediated by an initial depression of deprived-eye responses followed by an increased responsiveness to the nondeprived eye. It is not fully clear to what extent these 2 events are correlated and which are their physiological and molecular mediators. The extracellular synaptic environment plays an important role in regulating visual cortical plasticity. Matrix metalloproteinases (MMPs) are a family of activity-dependent zinc-dependent extracellular endopeptidases mediating extracellular matrix remodeling. We investigated the effects of MMP inhibition on OD plasticity in juvenile monocularly deprived rats. By using electrophysiological recordings, we found that MMP inhibition selectively prevented the potentiation of neuronal responses to nondeprived-eye stimulation occurring after 7 days of MD and potentiation of deprived-eye responses occurring after eye reopening. Three days of MD only resulted in a depression of deprived-eye responses insensitive to MMP inhibition. MMP inhibition did not influence homeostatic plasticity tested in the monocular cortex but significantly prevented an increase in dendritic spine density present after 7 days MD in layer II–III pyramids.
The maturation of the visual function in response to external stimuli is one of the most studied examples of experience-dependent plasticity in the central nervous system (CNS). A proper vision crucially depends on the inputs received in early life: if one eye is defective during infancy, its visual acuity does not develop normally and binocularity is impaired, a pathological condition called amblyopia (Timney 1983; Kiorpes et al. 1998; Maurer et al. 1999; Prusky et al. 2000; Prusky and Douglas 2003). At cortical level, studies employing monocular deprivation (MD); i.e., the closure of one eye through eyelid suture) in animals have shown that visual cortical neurons become strongly dominated by the nondeprived eye and the proportion of binocular neurons greatly decreases (Hubel and Wiesel 1963; Berardi et al. 2000). A 2-step mechanism is responsible for this result: an early depression of the neuronal responses to the deprived-eye stimulation is followed by a delayed potentiation of the nondeprived-eye inputs (Frenkel and Bear 2004).
Despite intense efforts, it is not fully clear to what extent these 2 events are correlated and which are their physiological and molecular mediators. It has been demonstrated that brief MD could lead to the weakening of intracortical synaptic connections serving the closed eye through homosynaptic long-term depression (LTD) on excitatory synapses (Rittenhouse et al. 1999; Heynen et al. 2003; Frenkel and Bear 2004) or long-term potentiation (LTP) of local inhibitory transmission (Maffei et al. 2006). On the contrary, the literature on the mechanisms responsible for open-eye potentiation in juvenile mice remains scarce.
Both synaptic scaling (Mrsic-Flogel et al. 2007; Kaneko et al. 2008) and Hebbian plasticity (Smith et al. 2009) have been suggested to play a fundamental role in this context. The identification of specific molecular mediators is crucial to establish the real contribution of these 2 phenomena. Recent experimental data indicate that glial tumor necrosis factor (TNF)-α is required for synaptic scaling subserving potentiation of nondeprived response (Kaneko et al. 2008). On the contrary, a well-defined molecular pathway linking the potentiation of the open-eye responses with LTP mechanisms has still to be established (Smith et al. 2009).
Metalloproteinases (MMPs) are a family of zinc-dependent endopeptidases (Birkedal-Hansen 1995) that comprises more than 25 members divided into 4 major categories: collagenases, gelatinases, membrane type, and stromelysins (Wright and Harding 2009). MMPs contain a propeptide sequence that needs to be cleaved in order to unmask their catalytic domain. During CNS development, MMPs regulate proliferation, migration, differentiation and survival of various cell types, as well as axonal growth and myelination (Agrawal et al. 2008). In adult rats, hippocampal MMP levels increase transiently during Morris water maze learning, while inhibition of their activity prevents acquisition of the task and alters LTP (Meighan et al. 2006). MMP-9 activity increases selectively in the areas of the dentate gyrus undergoing dendritic remodeling in response to kainate injections (Szklarczyk et al. 2002) and after stimuli that induce late-phase LTP in the CA1 area of the hippocampus (Nagy et al. 2006). Conversely, LTP is impaired in brain slices from MMP-9 null-mutant mice (Nagy et al. 2006).
Taken together, these reports provide evidence for a crucial physiological role of MMPs in the normal functioning of the CNS and in the regulation of synaptic plasticity. In order to achieve a broader comprehension of MMP function and to investigate the molecular bases of the input-specific effects of MD, we analyzed the influence of MMP inhibition in MD rats during the critical period. We found that MMP activity is selectively essential for the potentiation of neuronal responses both after MD and during recovery of function once binocular vision (BV) is restored. MMP inhibition did not influence homeostatic plasticity but significantly prevented an increase in dendritic spine density present after 7 days MD in layer II–III pyramidal cells.
Materials and Methods
All animal work has been conducted according to relevant national and international guidelines. A total of 82 Long–Evans hooded rats aged between P21 and P45 were used in this study. Animals were separated from their mothers at P20 and were housed 2 or 3 per cage at 20–22 °C under a 12-h light/dark cycle and had ad libitum access to water and food. For the MD experiments, MD and minipump implant were performed under Avertin (tribromoethanol in amylene hydrate; 1 mL/hg of body weight, intraperitoneally) anesthesia at P21. Minipumps (model 1007D; rate 0.5 μL/h; Alzet) were connected to a cannula (gauge 30) implanted in the visual cortex contralateral to the deprived eye (4 mm lateral to the midline, 4 mm anterior to lambda, 3–4 mm anterior to the recording zone). Vehicle or GM6001 solution was continuously infused for 1 week in the visual cortex contralateral to the deprived eye. Animals were deprived contextually with minipump implant and recorded either after 1 week (7 days MD) or after 3 days of MD. Experiments of extracellular recordings in the monocular visual cortex were performed as described except for the stereotaxical coordinates of the cannula implant that were 3 mm lateral to the midline, 4 mm anterior to lambda, 3–4 mm anterior to the recording zone. In a subset of animals, 7 days MD was followed by eye reopening to allow a normal BV and the binocular region of the visual cortex was infused either with vehicle or with GM6001 solution for 2 weeks via osmotic minipump (model 2002; rate 0.5 μL/h; Alzet). After this period, extracellular recordings were performed and binocularity was assessed.
Assessment of MMP Blockade by GM6001
In order to assess residual MMP activity after infusion of increasing concentrations of GM6001, we performed an enzymatic assay (SensoLyte 520 Generic MMP Assay Kit and SensoLyte 520 MMP-9 Assay Kit; AnaSpec) on cortical extracts. These kits use a 5-FAM/QXL520 fluorescence resonance energy transfer (FRET) peptide as an MMP substrate. In the intact FRET peptide, the fluorescence of 5-FAM is quenched by QXL520. Upon cleavage into 2 separate fragments by MMPs, the fluorescence of 5-FAM is recovered and can be monitored at excitation/emission wavelengths 490/520 nm.
Tissue samples were homogenized in assay buffer containing 0.1% (v/v) Triton-X 100 and then centrifuged for 15 min at 10 000 × g at 4 °C. The supernatants were collected and stored at −80 °C until use. Samples’ protein concentration was determined by Bio-Rad assay (Bio-Rad, Italy), compared with a bovine serum albumin based standard curve.
An increasing concentration of GM6001 was added to a constant quantity of protein extract and substrate. Measurement of fluorescence intensity, performed using StepOne (Applied Biosystems) real-time PCR machine, started immediately at Ex/Em = 490/520 nm and recorded continuously every minute for 1 h. Relative fluorescence units were plotted versus time, the range of time points during which the reaction is linear was determined, and the slope of the linear portion of the data plot was calculated. The fluorescence reading from a well that did not contain GM6001 served as a control in each experimental session.
In vivo Electrophysiology
Extracellular recordings of single-unit activity were performed essentially as previously described (Lodovichi et al. 2000; Di Cristo et al. 2001; Pizzorusso et al. 2002). Recordings were performed under urethane anesthesia (0.7 mL/hg, i.p.; 20% solution in saline). For each animal, 7–10 cells were recorded in each of at least 2 tracks spaced evenly (200 μm) across the binocular or monocular primary visual cortex (Oc1B or Oc1M) contralateral to the deprived eye to avoid sampling bias. Only cells with receptive fields (RFs) within 20° of the vertical meridian were included in our sample. Spontaneous activity, peak response, responsiveness, and RF size were determined from peristimulus time histograms (PSTHs) recorded in response to computer-generated bars, averaged over at least 15 stimulus presentations (Lodovichi et al. 2000). Peak response and RF size were assessed using optimally oriented bars; responsiveness was calculated as the signal-to-noise ratio, calculating the ratio between peak response and spontaneous activity. Ocular dominance (OD) was quantitatively evaluated from PSTH according to the classification of Hubel and Wiesel (Hubel and Wiesel 1970), and contralateral bias index (CBI) was calculated as follows:
Dendritic Spine Analysis
Spine density measurements were performed by using the protocol of Mataga et al. (2004), with minor modifications. A total of 17 rats were used: 6 rats as control normal subjects, 6 were MD and treated with vehicle solution, 5 animals were MD and treated with GM6001. The animals were lightly perfused with 4% paraformaldehyde (40 mL, 6 mL/min). After decapitation, the brain was removed and immersed in ice-cold phosphate buffered saline (PBS) solution. A block of visual cortex was sectioned in the coronal plane into 300-μm-thick slices by using a vibratome (Leica, Vienna, Austria). The slices were transferred to a storage box containing PBS and maintained at room temperature. Dil lipophilic dye (1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine; Molecular Probes) was coated onto tungsten particles (diameter 1.3 μm; Bio-Rad) according to Gan et al. (2000); Dil coated particles were delivered to the slices by using a Helios Gene Gun System (Bio-Rad). A polycarbonate filter with a 3.0-μm pore size (Molecular Probes) was inserted between the gun and the preparation on a platform to remove clusters of large particles. Density of labeling was controlled by gas pressure (80 psi of helium). After labeling, slices were additionally post-fixed in 4% paraformaldehyde for 2 h. A Fluoview (Olympus, Tokyo) confocal microscope (×60 water immersion objective; numerical aperture 0.9) was used to image the labeled structures (×1.5 zoom). At least 5–15 labeled typical pyramidal neurons were randomly selected from layer II–III in the binocular zone of the visual cortex contralateral to the deprived eye. Images of basal and apical dendrites were acquired, stacked (0.5-μm step), and then analyzed with Metamorph (Molecular Devices) software blind to the treatment. Spine density and neck length were analyzed along each dendrite until 100 μm of distance from the soma, and then dendritic protrusions were classified by morphology in filopodia or spines.
Inhibition of MMPs during the Critical Period Partially Prevents the OD Shift Induced by MD in the Rat Visual Cortex
In order to evaluate the involvement of MMPs in visual cortex plasticity, we employed GM6001, a widely used broad-spectrum MMP inhibitor that has been proven to affect late-phase LTP in the CA1 field of the hippocampus (Nagy et al. 2006). The visual cortex was infused by means of a cannula connected to an osmotic minipump (Pizzorusso et al. 1999) containing either vehicle or GM6001 solution. Cannula implant was performed together with MD in P21 rats (see Materials and Methods).
After 7 days of MD and drug infusion, we performed extracellular recordings of neuronal spiking activity in the binocular region of the primary visual cortex. No differences in the mean spontaneous activity, peak response, responsiveness, or RF size (see Supplementary Fig. S1) was found in the GM6001 or vehicle groups with respect to normal rats indicating that the general neuronal properties of rats implanted with the osmotic minipump were preserved.
OD plasticity was evaluated using 2 different indexes: The CBI was used to estimate the strength of the responses to visual stimulation of the contralateral eye for each animal, and the cumulative distribution of the OD score of each neuron was used to assess the OD distribution of cortical cells. As shown in the Figure 1a, MD significantly reduced CBI in the vehicle-treated animals (mean CBI in saline-treated rats = 0.31 ± 0.03, n = 7; mean CBI in normal animals = 0.63 ± 0.04, n = 7), whereas in the GM6001 group, only a partial decrease was present (mean CBI in GM6001-treated rats = 0.43 ± 0.08, n = 5). The partial OD shift present in the GM6001-treated rats was even more evident computing the OD score distribution (Fig. 1b). Also using this index the OD shift caused by MD in vehicle-treated rats resulted partially prevented by GM6001.
MMP Inhibition Selectively Prevents the Potentiation of the Open-Eye Response after MD
Studies performed in the visual cortex of young mice have shown that MD induces a rapid depression of the responses to the closed eye that is followed by a delayed potentiation of neuronal activity driven by the stimulation of the open eye (Frenkel and Bear 2004).
In order to study the contribution of each eye-specific pathway to the OD shift caused by MD, we evaluated the peak response of visual cortical neurons after presentation of a light bar either to the contralateral or to the ipsilateral eye. As expected, in normal animals, we found a significant contralateral prevalence in the control of neuronal activity (mean peak response of the contralateral eye = 16.81 ± 1.72 spikes/s; mean peak response of the ipsilateral eye = 12.85 ± 2.03 spikes/s; n = 7; Fig. 2a), while in the vehicle MD group, there was a significant decrease in the peak response driven by the contralateral deprived eye and an increase of the response to the nondeprived eye (mean peak response of the contralateral eye = 10.69 ± 1.22 spikes/s; mean peak response of the ipsilateral eye = 20.41 ± 2.62 spikes/s; n = 7; Fig. 2a). Interestingly, MMP inhibition prevented selectively the potentiation of the open-eye pathway and did not influence the depression of the contralateral deprived eye (mean peak response of the contralateral eye = 11.41 ± 0.70 spikes/s; mean peak response of the ipsilateral eye = 11.36 ± 0.90 spikes/s; n = 5; Fig. 2a).
Thus, only one of the mechanisms underlying MD effect was affected by MMP inhibition explaining the partial OD shift observed through CBI and OD score analysis in MMP inhibitor–treated rats.
To rule out the possibility that the partial effect of GM6001 on OD plasticity was due to a low GM6001 dosage, we treated a group of MD rats with a higher GM6001 concentration (2 mM). The results were indistinguishable from that obtained with the low GM6001 dosage. Indeed, also with this high GM6001 dose, a specific block of potentiation of nondeprived-eye responses with no effect on deprived-eye depression was observed (mean peak response in the contralateral eye = 10.37 ± 1.27 spikes/s; mean peak response in the ipsilateral eye = 10.69 ± 1.20 spikes/s; n = 6; Fig. 2a).
The cumulative distributions of cells’ peak response values confirmed these data revealing a significant reduction of the peak response to the contralateral eye stimulation in saline-treated and in both doses of GM6001-treated rats with respect to normal animals (Fig. 2b). Moreover, potentiation of the ipsilateral nondeprived-eye responses was evident only in the distribution of the saline group (Fig. 2c).
Blockade of gelatinase activity was complete for GM6001 concentrations higher than 10 nM (Fig. 3a). Noteworthy, among gelatinases, only MMP-9 expression was regulated by MD. Indeed, MMP-9 mRNA was found to be significantly reduced in the visual cortex contralateral to the deprived eye in MD animals (Fig. 3b), while MMP-2 and TIMP-1 levels remained unchanged (Fig. 3b).
Short-term MD Determines a Decrease in the Response to the Deprived Eye Both in Vehicle- and GM6001-Treated Rats
In order to confirm that MMP inhibition only affected the potentiation of the open-eye responses after MD, visual deprivation was performed for a shorter period (3 days). It has been suggested that a short period of MD (2–3 days) in mice determines a reduction of deprived-eye responses while not influencing the activity of the nondeprived-eye pathway (Frenkel and Bear 2004). Therefore, if the effect of MMP inhibition was specific on potentiation of nondeprived-eye responses, no influence of GM6001 on short-term MD should be expected. To test this hypothesis, we analyzed neuronal responses in rats treated with GM6001 or vehicle after 3 days of MD. We found that a significant decrease of the peak response to the stimulation of the deprived eye was present in both groups of treatment (mean contralateral peak response in saline-treated rats = 10.47 ± 1.00 spikes/s, Sal, n = 5; mean contralateral peak response in GM6001-treated rats = 9.09 ± 0.37 spikes/s, GM6001, n = 6). Conversely, no sign of potentiation was evident in the responses of the nondeprived eye both in saline- and GM6001-treated rats (mean ipsilateral peak response in saline-treated rats = 12.08 ± 1.30 spikes/s; mean ipsilateral peak response in GM6001-treated rats = 11.35 ± 0.62 spikes/s; Fig. 4a). Once again, the analysis of cumulative distributions of peak responses for each group confirmed these results (Fig. 4b,c). A summary of the results obtained after MD from both groups at different time points and after the stimulation of either eye is represented in Figure 5a.
Potentiation of Deprived-Eye Responses after Eye Reopening Is Blocked by MMP Inhibition
If MMPs are required for experience-dependent potentiation of visual responses, their inhibition should also block the recovery of the responses occurring when the deprived eye is reopened in juvenile rats. To test this possibility, the deprived eye was reopened after 7 days of MD and rats could experience BV for 2 weeks prior to OD assessment. During this period, GM6001 or vehicle was infused into the visual cortex contralateral to the formerly deprived eye. OD recordings revealed that in saline-treated animals, there was a significant recovery of deprived-eye response after 2 weeks of BV (mean contralateral peak response in saline-treated rats = 16.09 ± 1.34 spikes/s, n = 7; Fig. 5b). Treatment with GM6001 prevented this effect (mean contralateral peak response in GM6001-treated rats = 12.93 ± 1.62 spikes/s, n = 6; Fig. 5b). On the contrary, peak responses of the nondeprived eye returned to normal levels both in saline- (mean ipsilateral peak response in saline-treated rats = 13.5 ± 0.82 spikes/s, n = 7) and GM6001-treated animals (mean ipsilateral peak response in GM6001-treated rats = 12.19 ± 0.85 spikes/s, n = 6; Fig. 5b).
MMP Inhibition Does Not Affect Homeostatic Plasticity in the Monocular Region of the Primary Visual Cortex
Hebbian mechanisms of synaptic plasticity such as LTP and LTD have been widely hypothesized to underlie numerous forms of experience-dependent plasticity in the brain, including learning, memory, and activity-dependent development (Shah and Crair 2008). Several papers support the idea that the loss and gain of eye-specific responses after MD might arise from these phenomena (Rittenhouse et al. 1999; Heynen et al. 2003; Frenkel and Bear 2004; Maffei et al. 2006). Nevertheless, a different form of non-Hebbian plasticity referred to as homeostatic plasticity has been observed in the visual cortex of MD animals (Maffei et al. 2004; Mrsic-Flogel et al. 2007; Kaneko et al. 2008).
In order to verify the involvement of MMPs in the phenomena of homeostatic plasticity in the visual cortex, we analyzed the response of neurons in the monocular region after 3 and 7 days of MD. In accordance with previous studies (Mrsic-Flogel et al. 2007), we found in normal deprived rats an initial reduction of neuronal responses after brief MD (mean peak response in normal nondeprived animals at p21 = 12.98 ± 0.82 spikes/s; mean peak response at 3 days of MD = 10.37 ± 0.19 spikes/s; n = 3) followed by an enhancement of visual responses after 7 days of visual deprivation (mean peak response at 7 days of MD = 15.29 ± 0.64 spikes/s, n = 4). These results were superimposable to those obtained in the animals treated with GM6001 (mean peak response at 3 days of MD = 10.50 ± 0.41 spikes/s, n = 4; mean peak response at 7 days of MD = 15.85 ± 0.23 spikes/s, n = 3; Fig. 6) suggesting that MMP function is not involved in the homeostatic regulation of neural responsiveness after MD.
GM6001 Prevents Structural Remodeling of Dendritic Spines after MD
Experience-dependent changes in brain circuitry have both functional and structural components (Hofer et al. 2009; Coleman et al. 2010). The formation and elimination of synapses is believed to be one of the mechanisms underlying adaptive remodeling of neural circuits (Trachtenberg et al. 2002; Majewska et al. 2006). The vast majority of postsynaptic excitatory synapses in the visual cortex reside on dendritic spines whose structure and shape are highly dynamic (Fischer et al. 1998; Dunaevsky et al. 1999; Yasumatsu et al. 2008) and largely depend on neural activity (Fischer et al. 2000; Oray et al. 2006; Hofer et al. 2009). We analyzed whether dendritic spine density is affected by MD and whether these changes involve MMP activity. Dendritic spines were visualized in the binocular region of the primary visual cortex through diolistic labeling. Spine density was significantly increased in the basal dendrites of II–III layer pyramidal neurons in vehicle-treated animals after 7 days of MD (mean spine density in normal animals = 0.70 ± 0.07 spine/μm, n = 6; mean spine density in saline-treated animals = 0.97 ± 0.07 spine/μm, n = 6), whereas no difference was found in the corresponding apical dendrites (data not shown). The increment in spine density was prevented by GM6001 (mean spine density = 0.72 ± 0.06 spine/μm, n = 5; Fig. 7a).
Electron microscopy reconstruction revealed that spine appearance could be extremely heterogeneous including long and thin filopodia-like protrusions, short spines without a well-defined spine neck, and spine, with a large head (Harris and Stevens 1989; Harris and Kater 1994). In our study, the rise in spine density in the vehicle group was accompanied by an increase of filopodia-like structures percentage (mean filopodia percentage in normal animals = 10.39 ± 1.49%; mean filopodia percentage in saline-treated animals = 16.98 ± 2.78%; Fig. 7b) documented also by an overall shift of spine length distribution toward higher values (mean spine length in normal animals = 1.32 ± 0.04 μm, n = 239 spines; mean spine length in saline-treated animals = 1.47 ± 0.03 μm, n = 566 spines; Fig. 7c). Once again, the treatment with the MMP inhibitor prevented these effects (mean filopodia percentage in GM6001-treated animals = 12.24 ± 1.93%; mean spine length in GM6001-treated animals = 1.27 ± 0.04 μm; n = 385 spines; Fig. 7b,c). Thus, MMP inhibition prevented experience-dependent increase in spine density in II–III layer pyramidal cells in the visual cortex of MD rats.
Visual cortex plasticity is one of the most studied examples of experience-dependent alteration of brain structure and function. Decades of research, however, have not been sufficient to completely reveal the complexity of neuronal rearrangements underlying the OD shift caused by MD. Recent observations in mice have demonstrated that the closure of one eye during the critical period induces a rapid reduction in the visual neural responses to the closed-eye stimulation and a delayed increase of the responses to the open eye (Frenkel and Bear 2004; Kaneko et al. 2008). While confirming these results in the rat, we also found a distinct molecular pathway, triggered by MMP activation, whose functioning is fundamental for the experience-dependent potentiation of visual responses. Indeed, inhibition of MMP activity selectively prevented the increase in the response to the nondeprived-eye stimulation after 7 days of MD, whereas no effect was present on the depression of the responses to the deprived eye both after 3 or 7 days of MD. Furthermore, the recovery of the cortical responses to the deprived eye observed after a period of reestablished binocular visual experience was prevented by MMP inhibition suggesting that a common mechanism could be responsible both for the strengthening of active inputs during MD and for the reestablishment of a normal visual function once binocularity is restored.
Our data support the hypothesis that the bidirectional effects elicited by MD on deprived- and nondeprived-eye responses are mechanistically different. It has been suggested that short-term MD could lead to depression of deprived-eye responses through homosynaptic LTD involving changes in the phosphorylation state of 2-amino-3-(5-methyl-3-oxo-1,2- oxazol-4-yl)propanoic acid (AMPA) receptors and activation of CB1 receptors (Smith et al. 2009). Furthermore, recent data show that MD reduces the intrinsic excitability of layer V pyramidal neurons (Maffei et al. 2006). Conversely, the mechanisms underlying potentiation of the responses to the nondeprived eye remain to be fully clarified. Both homeostatic plasticity and LTP have been suggested to play a crucial role in this process. Recent data indicate that glial TNF-α, which is responsible for synaptic scaling after prolonged blockade of activity (Stellwagen and Malenka 2006), is also required for potentiation of nondeprived-eye response after MD (Kaneko et al. 2008). Evidence for homeostatic changes in neuronal activity have been documented in slices of deprived visual cortices (Kirkwood et al. 1996; Desai et al. 2002; Maffei et al. 2004; Philpot et al. 2007). Moreover, visual response homeostasis has been found to occur in the monocular neurons of intact animals after transient deprivation (Mrsic-Flogel et al. 2007). Our data indicate that MMPs do not participate in these kinds of events since homeostatic plasticity in the monocular region of the visual cortex after 7 days of MD is not impaired by their inhibition.
Interestingly, MMPs seem to have an important role in various forms of experience-dependent plasticity and LTP. In particular, MMP-9 is necessary for spatial and emotional learning (Meighan et al. 2006; Nagy et al. 2006) as well as for late-phase LTP in the hippocampus (Nagy et al. 2006) and prefrontal cortex (Okulski et al. 2007). We have demonstrated that, among different MMPs, gelatinases display a complete blockade of activity at both the concentrations of GM6001 used. However only MMP-9 expression was regulated by neuronal activity suggesting that this specific enzyme could be primarily involved in the phenomena described.
It has been shown that late-phase LTP is accompanied by changes in dendritic spine number and morphology (Yuste and Bonhoeffer 2001; Lynch et al. 2007). Dendritic spines receive the majority of excitatory inputs in the CNS, and recent in vivo results have shown that their turnover, shape, and motility are influenced by sensory experience (Trachtenberg et al. 2002; Mataga et al. 2004; Majewska et al. 2006; Hofer et al. 2009). Short-term protocols of MD induce a transient decrease in spine density in the binocular region of the visual cortex (Mataga et al. 2004), likely reflecting the early depression of deprived-eye inputs. In the same area, after 7 days of MD, we observed a significant increase of spine density in the basal dendrites of layer II–III pyramidal cells in control MD rats. This result may provide a possible anatomical explanation for the delayed enhancement of the open-eye responses following MD. Accordingly, MMP inhibition not only prevented the potentiation of nondeprived inputs but also affected the rise in the number of spines. These data parallel recent observations showing that spine number is enhanced by 8 days of MD also in the binocular visual cortex of adult mice (Hofer et al. 2009) even if this effect was observed only in layer V pyramidal cells.
In summary, we propose a model in which Hebbian and homeostatic processes may act in concert to produce the increase in visual response to the nondeprived eye after MD. MMPs could be essential molecular mediators of the experience-dependent potentiation process and could influence structural remodeling, the final consolidating step of the entire plasticity event.
MMPs could regulate structural plasticity by interfering with the signaling of synapse-associated molecules (Michaluk et al. 2007; Michaluk et al. 2009) and/or cleaving those components of the extracellular matrix that hamper structural modifications (Zuo et al. 1998). MMPs are activated by treatments that result in sustained neuronal activity (Wilczynski et al. 2008; Takacs et al. 2010) and LTP (Nagy et al. 2006). Furthermore, MMPs are expressed in dendritic spines (Konopacki et al. 2007) in colocalization with excitatory synaptic markers (Gawlak et al. 2009). Our data suggest that MMPs might be particularly important in the process of spinogenesis since their inhibition prevented the increase in the percentage of filopodia-like structures occurring in 7-day MD rats.
MMPs’ role in OD plasticity is more selective than that of other proteinases like tissue plasminogen activator (tPA). Indeed, knockout mice for tPA display a complete blockade of dendritic spine pruning after MD and no OD plasticity during the critical period (Mataga et al. 2002). By contrast, MMP inhibition only affected the potentiation of nondeprived-eye responses. It could be that tPA and MMPs act on plasticity through completely different mechanisms; however, MMPs have been reported to be a substrate for tPA (Wang et al. 2003; Tsuji et al. 2005; Suzuki et al. 2007) and could, therefore, contribute in mediating part of its actions.
MMPs constitute a large family with more than 20 members (Yong et al. 2001). Their biochemical complexity and the large variety of substrates account for the richness of their functions. An increase in MMP activity is generally considered detrimental in several diseases of the CNS such as multiple sclerosis, HIV, stroke, and spinal chord injury (Agrawal et al. 2008) making MMP inhibitors attractive therapeutical agents for these pathologies (Leppert et al. 2001; Dang et al. 2008; Leonardo et al. 2009; Liu et al. 2009). However, it has been found that during a delayed phase after stroke, an increase of MMP activity in perilesional cortical areas may contribute to the plastic remodeling of circuits (Zhao et al. 2006; Sood et al. 2008) promoting anatomical and functional recovery. Therefore, inhibition of MMPs could have negative effects when mechanisms of neural plasticity are particularly needed. Our data support this possibility showing that MMPs are necessary for specific forms of experience-dependent plasticity.
MIUR-PRIN grant; European Union 7th Framework Programme (FP2007-2013) EUROV1SION and PLASTICISE (223326, 223524); EXTRAPLAST IIT project; EU COST Action BM1001 ECMNET.
Conflict of Interest: None declared.