Matrix Metalloprotease 3 Activity Supports Hippocampal EPSP-to-Spike Plasticity Following Patterned Neuronal Activity via the Regulation of NMDAR Function and Calcium Flux

Abstract

Matrix metalloproteases (MMPs) comprise a family of endopeptidases that are involved in remodeling the extracellular matrix and play a critical role in learning and memory. At least 24 different MMP subtypes have been identified in the human brain, but less is known about the subtype-specific actions of MMP on neuronal plasticity. The long-term potentiation (LTP) of excitatory synaptic transmission and scaling of dendritic and somatic neuronal excitability are considered substrates of memory storage. We previously found that MMP-3 and MMP-2/9 may be differentially involved in shaping the induction and expression of excitatory postsynaptic potential (EPSP)-to-spike (E-S) potentiation in hippocampal brain slices. MMP-3 and MMP-2/9 proteolysis was previously shown to affect the integrity or mobility of synaptic N-methyl-D-aspartate receptors (NMDARs) in vitro. However, the functional outcome of such MMP-NMDAR interactions remains largely unknown. The present study investigated the role of these MMP subtypes in E-S plasticity and NMDAR function in mouse hippocampal acute brain slices. The temporal requirement for MMP-3/NMDAR activity in E-S potentiation within the CA1 field largely overlapped, and MMP-3 but not MMP-2/9 activity was crucial for the gain-of-function of NMDARs following LTP induction. Functional changes in E-S plasticity following MMP-3 inhibition largely correlated with the expression of cFos protein, a marker of activity-related gene transcription. Recombinant MMP-3 promoted a gain in NMDAR-mediated field potentials and somatodendritic Ca²⁺ waves. These results suggest that long-term hippocampal E-S potentiation requires transient MMP-3 activity that promotes NMDAR-mediated postsynaptic Ca²⁺ entry that is vital for the activation of downstream signaling cascades and gene transcription.

Keywords: E-S potentiation; Extracellular proteolysis; Hippocampus; Matrix metalloprotease; NMDAR; Synaptic plasticity.

Fig. 1

Temporal dependence of hippocampal E-S potentiation on NMDARs and MMP-3 activity. a Recording scheme. Two recording (REC) electrodes simultaneously monitored fEPSPs and population spikes in the CA1 region in response to Schaffer collateral stimulation (STIM). b Left, Time-course of maximal fEPSP slopes normalized to baseline values in control slices (black circles) and when bath-applied with the NMDAR antagonist APV (50 μM) at varying time points relative to HFS: before HFS (red triangles), 15 min post-HFS (green squares), or 60 min post-HFS (blue diamonds). The top shows exemplary traces of fEPSPs before HFS (1) and 90 min after HFS (colors match figure legend). Scale bar = 0.5 mV, 20 ms. Right, Time-course of population spike (PS) amplitudes normalized to baseline values in control slices (black circles) and when bath-applied with the NMDAR antagonist APV (50 μM) at varying time points relative to HFS: before HFS (red triangles), 15 min post-HFS (green squares), or 60 min post-HFS (blue diamonds). The top shows exemplary and normalized traces of PS before HFS (1) and 90 min after HFS (colors match figure legend). Scale bar = 0.5 mV, 10 ms. c Statistics of average PS/fEPSP ratio presented in b at 90 min post-HFS. The asterisk indicates a significant difference vs. slices in which HFS was applied in the absence of APV. Notice that APV that was applied up to 30 min post-HFS (green) significantly attenuated the PS/fEPSP upregulation following HFS. d Impact of APV on E-S coupling before HFS (circles) and 90 min after HFS (triangles). Relationships between stimulus strength and the PS/fEPSP slope ratio are shown. Notice that APV that was applied up to 30 min post-HFS (yellow) significantly attenuated the PS/fEPSP upregulation following HFS that was observed in CTR (black). e Effect of MMP-3 inhibitor NNGH (10 μM) on E-S potentiation. Middle and right, fEPSP slope and population spike amplitude time-course, respectively, as in b, except instead of APV, an MMP-3 inhibitor was applied at varying time points with regards to HFS. Left, Exemplary traces of fEPSPs (top) and population spikes (bottom) before and 90 min after HFS in the presence of NNGH (colors match figure legends). Scale bars as in b. f Statistics of average PS/fEPSP ratio presented in e at 90 min post-HFS. The asterisk indicates a significant difference vs. slices in which HFS was applied in the absence of NNGH. Notice that NNGH that was applied up to 15 min post-HFS (green) significantly attenuated the PS/fEPSP upregulation following HFS. g Impact of NNGH on E-S coupling before HFS (circles) and 90 min after HFS (triangles). Relationships between stimulus strength and the PS/fEPSP slope ratio are shown. Notice that NNGH that was applied up to 15 min post-HFS (green) significantly attenuated the PS/fEPSP upregulation following HFS that was observed in CTR (black). The zero value on the time bars represents the moment of tetanization (HFS, 4 × 100 Hz). The horizontal colored bars represent drug application. The numbers on the graphs refer to the number of experiments. *p < 0.05

Fig. 2

Synaptic NMDAR potentiation critically depends on MMP-3 and integrin signaling but not MMP-2/9. (a) Isolation of synaptic NMDAR-mediated fEPSPs (fEPSP_NMDA). Upper, Sample traces of fEPSPs recorded in Mg²⁺-free solution before (gray) and after (black) application of the AMPA/kainate antagonist DNQX (20 μM) and l-type channel blocker nifedipine (20 μM). Lower, fEPSP_NMDA (gray) was completely abolished upon bath application of APV (50 μM) (black). (b ₁) Exemplary traces of fEPSP_NMDA before HFS (1) and 90 min after HFS (colors match (b ₂) legend). Scale bar = 0.2 mV, 20 ms. (b ₂) Upper, Time-course of synaptic NMDAR-mediated fEPSP under control conditions (black) and upon bath application of MMP-3 inhibitor before HFS (red), 15 min post-HFS (green), or 30 min post-HFS (yellow). Lower, Time-course of synaptic NMDAR-mediated fEPSP under control conditions (black) and upon bath application of drugs before HFS: MMP-3 inhibitor UK356618 (2 μM) (magenta), MMP-2/9 inhibitor (10 μM) (cyan), or integrin-blocking peptide GRGDSP (30 min post-HFS) (gray). The zero value on the time bars represents the moment of tetanization (HFS, 4 × 100 Hz). The horizontal colored bars represent drug application. (c) Statistics of effects of drugs shown in (b ₂) measured 60 min post-HFS. The numbers on the graphs refer to the number of experiments. *p < 0.05

Fig. 3

MMP-3 or NMDAR inhibition affects c-Fos expression following patterned neuronal activity. a Exemplary confocal immunofluorescence images of CA1 hippocampal region (×40 magnification) in acute brain slices fixed immediately following cessation of the electrophysiological recordings. Basally stimulated slices (0.1 Hz) (upper panels) were compared with slices fixed 90 min post-E-S potentiation induction in the absence (middle panel) or presence (lower panel) of APV. The sections were stained against cFos (green channel) and NeuN (red channel), and the colocalization of both markers was analyzed (merged channel). b Statistics of the effects of electrical stimulation and drug treatment on cFos expression in the CA1 neuronal population stained with NeuN. Slices were basally stimulated (0.1 Hz) (gray), or E-S potentiation was induced with HFS (4 × 100 Hz) in the absence (LTP) (black) or presence of the NMDAR antagonist APV applied before or 15 and 30 min after HFS. c Exemplary immunofluorescence images of CA1 sections following the induction of synaptic LTP_NMDA (see Fig. 2). Basally stimulated slices (0.1 Hz) (upper panels) were compared with slices that were fixed 90 min post-LTP_NMDA induction with HFS in the absence (middle panel) or presence (lower panel) of the MMP-3 inhibitor NNGH. See a for channel description. d Statistics of the effects of electrical stimulation and NNGH treatment on cFos expression in CA1 neurons following recordings of synaptic NMDAR-mediated responses (shown in Fig 2). Slices were basally stimulated (0.1 Hz) (gray), or LTP_NMDA was induced with HFS (4 × 100 Hz) in the absence (LTP) (black) or presence of the MMP-3 inhibitor NNGH applied before or 15 and 30 min after HFS. The numbers on the graphs refer to the number of sections analyzed. *p < 0.05 vs. basally stimulated slices

Fig. 4

Recombinant MMP-3 promotes NMDAR-mediated responses and somatodendritic Ca²⁺ waves following multiple exposures to NMDA. a Recording scheme for glutamate-evoked NMDAR responses in acute hippocampal brain slice in the presence of inhibitor cocktail and Mg²⁺-free aCSF. A recording (REC) electrode recorded NMDAR-mediated field potentials in the CA1 stratum radiatum in response to local pressure injection of glutamate. b Representative NMDAR-mediated field potentials following short (200 ms) pressure injections of an external solution that contained the inhibitor cocktail (aCSF) (upper trace), aCSF + glutamate + d-serine (100 μM) without the NMDAR antagonist APV (middle trace), or aCSF + glutamate + d-serine with APV (lower trace). c Time-course of APV-sensitive field potential amplitude in response to glutamate (every 2 min) in the absence (black circles) or presence of recombinant MMP-3 (1 μg/ml). Notice the slowly emerging potentiation of field potential amplitude with time. d, e Exemplary differential images of relative Fura2 fluorescence change in somatodendritic compartment (pink) before (d) and during NMDA application (e, 60 μM; excitation wavelength, 340/380 nm; emission wavelength, 510 nm). The recording solution was supplemented with a cocktail of inhibitors to ensure NMDAR activation (see “Materials and Methods” section for details). f Exemplary time-course of Fura2 fluorescence (ΔF_340/380) following exogenous application of NMDA (six applications every 5–6 min) in control neurons (black trace), neurons incubated with the MMP-3 inhibitor NNGH (10 μM) (gray trace), and neurons pretreated with recombinant MMP-3 (1 μg/ml) for 30 min before recordings (red trace). Notice that MMP-3 pretreatment typically enhanced the amplitude of NMDA-evoked Ca²⁺ waves, whereas MMP-3 inhibition reduced the amplitude of NMDA-evoked Ca²⁺ waves. g Statistics of data shown in f for all KCl-sensitive cells. The cells were classified as undergoing potentiation (>105 %) (green), no change (95 > 105 %) (yellow), or depression (<95 %) (brown) of NMDA-evoked Ca²⁺ wave amplitudes compared with first NMDA application. Notice that pretreatment with recombinant MMP-3 (n = 171 neurons, n = 3 cultures) (middle panel) increased the fraction of neurons that underwent Ca²⁺ wave amplitude potentiation with time (CTR, n = 408 neurons, n = 5 cultures; χ ²: p < 0.001, sixth vs. first) (upper panel). The MMP-3 inhibitor NNGH had the opposite effect (n = 395 neurons, n = 4 cultures) (lower panel). *p < 0.05 vs. control slices (sixth to first). n number of slices. *p < 0.05 vs. control slices