The Journal of Neuroscience, February 15, 2006, ():
Matrix Metalloproteinase-9 Is Required for Hippocampal Late-Phase Long-Term Potentiation and Memory
J. Neurosci. Nagy et al.
26: 1923
Supplemental data
Files in this Data Supplement:
- supplemental material -
Supplemental references
- supplemental material
-
Figure S1 MMP inhibitors used in the study block MMP-2 and –9. A fluorometric enzymatic assay was used to verify potency and to confirm that the commercial MMP inhibitors used inhibited MMP-2 and –9 proteolytic activity. Recombinant active MMP-9 (A, C) or MMP-2 (B, D) (Calbiochem; 3.3 ng/ml) was added to DQ gelatin (3.3 µg/mL) and a range of concentrations of GM6001 (top graphs; in nM: 0.005, 0.05, 0.5; in µM: 1, 25) or Inhibitor II (bottom graphs; in nM: 0.17, 1.7, 17, 100; in µM: 25, 50). Aliquots of each mixture were incubated at room temberature for 2 hrs in assay buffer (EnzChek Gelatinase/Collagenase Assay Kit, Molecular Probes). Emitted fluorescence was measured, normalized to background (DQ-gelatin alone) and repeated at least twice in quadruplicates. Inhibition curves were generated with Prizm GraphPad software. IC50 values for both inhibitors were comparable to reported values.
- supplemental material
-
Figure S2 Specificity of MMP antisera. Whole rat hippocampal lysates (whole lysate) were loaded alongside 10 ng human recombinant pro- (inactive) and active-forms of MMP-2 and MMP-9. Membranes were immunoblotted with rabbit polyclonal antisera raised against active MMP-9 (A) or –2 (B) (1:500, Torrey Pines). Each antiserum recognized the appropriate recombinant pro- and active-forms of the MMP against which it was raised and did not cross-react with any form of the inappropriate MMP. In the case of MMP-2, additional bands were detected in hippocampal lysates that have been attributed previously to autocatalytic cleavage products (Davis and Martin, 1990; Sternlicht and Werb, 2001).
- supplemental material
-
Figure S3 Levels and activity of MMP-9 do not fluctuate in response to tissue slicing. Slicing hippocampal tissue can disrupt cellular integrity (Fiala et al., 2003), and levels of certain proteins in hippocampal slices have been shown previously to fluctuate simply as a result of such slicing (Taubenfeld et al., 2002). To rule out a confounding possibility that changes in levels of the active form of MMP-9 partly reflect an “injury” response to tissue slicing, we compared protein and activity levels of MMP–9 between slices frozen immediately after dissection with ones incubated in oxygenated Ringer’s solution for variable periods up to 150 min.
(A) Representative immunoblot of homogenates of hippocampal slices (n=3); membranes were probed with an antibody that recognizes the active form of MMP-9 (act-9) or one that recognizes tubulin (tub.). The blot shows that there are no fluctuations in levels of the active-form of MMP–9 across time as a result of tissue slicing.
(B) Representative in vitro gelatin zymograph of hippocampal slice homogenates showing that enzymatic activities of pro- and active-forms of MMP-9 are not fluctuating over time in response to tissue slicing.
- supplemental material
-
Figure S4 Basic properties of synaptic neurotransmission are unaffected in MMP-9 null-mutant mice.
(A) There are no differences between genotypes in synaptic input–output curves elicited in hippocampal slices from MMP-9 KO (black circles) or wildtype (wt) mice (white circles). Elicited fEPSP slopes (mV/ms) as a function of stimulation strength (V) are shown (n = 6 slices, 3-4 mice per genotype).
(B) There are no differences in facilitation induced by paired stimuli (PPF) with interpulse intervals (IPI) between 50-200 ms in slices from MMP-9 KO (black circles) and wt mice (white circles) (n = 6 slices, 3-4 mice per genotype).
- supplemental material
-
Figure S5 Baseline freezing in MMP-9-/- and wild type mice.
(A) There is no difference in context baseline freezing prior to training between wild type (+/+, black bar; n = 27) and MMP-9-/- mutant mice (-/-, white bar, n = 25). F(1,50) = 0.244, p = 0.6234.
(B) During the cued test, MMP-9-/- mutant mice (n = 16) show decreased baseline freezing in context “B” prior to the tone compared to wild type mice (n = 19). Although significantly different, both freeze less than 10% of the test time. F(1,33) = 5.760, *p = 0.022.
- supplemental material
-
Figure S6 Pavlovian fear conditioning of MMP-9-/- and wild type mice using activity suppression as a memory measure.
(A) MMP-9-/- mutant mice (-/-, white bar, n = 25) show decreased activity suppression during the context test compared to wild type mice (+/+, black bar, n = 27). F(1,50) = 4.595, *p = 0.037.
(B) Wild type (n = 19) and MMP-9-/- mutant mice (n = 16) show the same level of activity suppression during the cued test (F(1,33) = 0.006, p = 0.94). Thus, the two measures of memory, activity suppression and freezing, show consistently that MMP-9 null mice experience a deficit in long-term hippocampus-dependent memory.
- supplemental material
-
Figure S7 Model of MMP-9 function in hippocampal synaptic physiology and plasticity.
A speculative model illustrating regulation and function of MMP-9 in L-LTP. In a “resting” hippocampal synapse (left), NMDA receptors (black bars) are closed, a pool of mostly pro- (inactive) MMP-9 is present extracellularly; certain extracellular matrix (ECM) proteins harboring latent RGD sequences are found in the neuropil. Upon L-LTP-inducing stimuli (shaded triangle, top), levels of the active form of MMP-9 increase and MMP-9 becomes proteolytically active (middle) by mechanisms that require at least in part, NMDA receptor activation (open bars) and protein synthesis. The intracellular signaling cascades linking NMDA receptor activation to protein synthesis (straight arrows, middle) are presently unknown (question mark). There are probably other mechanisms that can rapidly activate the preexisting pool of latent pro-MMP-9, but these are also unknown at present. In potentiated hippocampus, MMP-9 associates with glia, dendrites and synapses. Speculatively, active MMP-9 may cleave certain ECM substrates, exposing the latent RGD sequences (middle and right). This, in turn, leads to activation of RGD-binding integrin receptors (right), which initiates intracellular signal transduction cascades (curved arrows, right) leading to increases in AMPA and NMDA glutamate receptor currents (bold open bars) as shown previously by Gall, Lynch and colleagues (see text).
