Reducing hippocampal extracellular matrix reverses early memory deficits in a mouse model of Alzheimer’s disease

For details about APP/PS1 mice see The Jackson Laboratory (strain B6C3-Tg(APPswe,PSEN1dE9)85Dbo/J; stock number 004462; http://​jaxmice.​jax.​org/​). APP/PS1 transgenic mice express a chimeric mouse/human APP gene harboring the Swedish double mutation K595N/M596L (APPswe) and a human PS1 gene harboring the exon 9 deletion (PS1dE9), both under the control of the mouse prion protein promoter (MoPrP.Xho) [[14],[20],[21]]. Mice were maintained as hemizygous and crossed with wildtype C57BL/6 mice. All experiments were performed with male mice and approved by the animal ethics committee of either the Royal Netherlands Academy of Arts and Sciences or the VU University Amsterdam.

Details can be found in the Additional file 1. In short, synaptosomes were isolated from hippocampi of APP/PS1 mice and wildtype littermates at 1.5, 3, 6 and 12 months of age as described previously [[22],[23]]. Five 8-plex iTRAQ experiments (i.e., five biological replicates per age group per genotype) were performed. Samples were analyzed using an ABI 5800 proteomics analyzer (Applied Biosystems, Forster City, CA). Protein identification and quantification were performed as described [[24]]. Mascot (MatrixScience, version 2.3.01) searches were performed against Swissprot (version 20/10/2010) and NCBInr (version 20/10/2010) databases. Proteins were considered for quantification if at least three peptides were identified in three replicate iTRAQ sets and at least one peptide in all other sets. Protein abundance was determined by taking the average normalized standardized iTRAQ peak area of all unique peptides annotated to that protein. Statistical significance was determined by calculating permutation-derived false discovery rates (FDR) using the SAM [[25]] package in MeV (version 4.6.1) [[25],[26]]. Changes in protein expression were considered to be significant when the FDR is <10% and log-fold change >0.125.

The Functional Classification Tool in the Database for Annotation, Visualization and Integrated Discovery (DAVID; http://​david.​abcc.​ncifcrf.​gov/​) [[27],[28]] was used to group proteins based on functional similarity and to determine enrichment of functional protein groups within the total set of differentially expressed proteins. Enrichment was determined using the total set of all detected proteins as the background set, and using the following settings: Similarity Term Overlap: 4; Similarity Threshold: 0.35; Initial Group Membership: 4; Final Group Membership: 4; Multiple Linkage Threshold: 0.50. Significance was determined based on the Fisher’s exact p-values reported by DAVID.

Immunoblot analysis was performed on six independent synaptosome protein extracts. To facilitate detection of ECM proteins, protein pellets were treated with chondroitinase ABC (Sigma Aldrich, Zwijndrecht, The Netherlands) at 0.5 U/50 mg protein for 90 min at 37°C. Of each sample, 10 μg protein was mixed with SDS sample buffer and heated at 90°C for 5 min. Proteins were separated on a Criterion™ TGX Stain-Free Precast Gel (4-16% Tris-Glycine; Bio-Rad) in a Criterion™ Cell Electrophoresis System (Bio-Rad), and electroblotted onto PVDF membrane overnight at 4°C. After blocking with 5% (v/v) non-fat dry milk and 0.5% (v/v) Tween-20 in Tris-buffered saline (TBS) for 1 h at RT, blots were incubated with primary antibodies, followed by a horseradish peroxidase-conjugated secondary antibody (Dako, Glostrup, Denmark; 1:10000). The following antibodies were used: 6E10 (Signet; 1:15000), anti-Brevican (gift from Dr. C. Seidenbecher, Magdeburg, Germany; 1:2000), anti-Tenascin-R (P.Glia, Bonn, Germany; 1:2000), anti-Hapln1 (Abcam; 1:1000), anti-Neurocan (Sigma; 1:1000). Blots were incubated with ECL substrate (GE Healthcare, Pollards Wood, UK), scanned with a Odyssey Imager (LI-COR) and analyzed with Image Studio software (LI-COR, version 1.1.7) using background correction. To correct for differences in sample input, all gels were imaged before electroblotting and the total protein densitometric values were used for sample normalization, which is more reliable then normalizing to the levels of a single protein [[29],[30]]. Significance was determined using a Student’s t-test.

Immunohistochemistry on brain sections at 3, 6, 9 and 12 months of age was performed as previously described [[31]]. Brains were fixed by transcardial perfusion with 4% (v/v) paraformaldehyde in phosphate buffered saline (PBS), pH 7.0. Coronal cryosections (10 μm) were thaw-mounted on Superfrost Plus slides, dried for 1 h at RT and stored at −20°C until use. Sections were fixed with fresh 4% (v/v) paraformaldehyde in 0.1 M PBS, pH 7.0, for 10 min at RT. Sections were treated with 10 mM sodium citrate and 0.05% (v/v) Tween-20 (pH 6.0) for 20 min at 95°C, blocked with 10% (v/v) normal donkey serum and 0.4% (v/v) Triton X-100 in PBS, followed by incubation overnight with 6E10 (Signet; 1:15000) and anti-GFAP (DAKO; 1:2000) at RT. Antigens were visualized using Cy3- and DyLight488-labeled secondary antibodies (Jackson ImmunoResearch Laboratories and Invitrogen; 1:1400), incubated for 2 h at RT. Sections were washed and coverslipped in Vectashield including DAPI as a nuclear dye (Vector Laboratories). PV and WFA stainings were performed on free-floating brain sections obtained from animals at 3 months of age. Sections were quenched with 10% (v/v) methanol and 0.3% (v/v) H₂O₂ in PBS for 30 min at RT, blocked with 0.2% (v/v) Triton X-100 and 5% (v/v) fetal bovine serum in PBS, and incubated overnight with anti-PV (Swant, Marly, Switzerland; 1:1000) and fluorescein labeled WFA (Vector Laboratories; 1:400) at RT. PV staining was visualized using a Alex568-labeled secondary antibodies (Invitrogen; 1:400), incubated for 2 h at RT. Sections were washed and coverslipped in Vectashield including DAPI as a nuclear dye (Vector Laboratories). PV and WFA staining were quantified using ImageJ v1.48. PV- and WFA-positive cells were counted using image thresholding and automated particle analysis in ImageJ (v1.48).

Mice were anesthetized with avertin (1.2% (w/v)), 0.02 ml/g, intraperitoneal) and chronically implanted with double guide cannulas in the CA1 region of the dorsal hippocampus as previously described [[32]]. Buprenorphine (0.1 mg/kg, subcutaneous) was injected as an analgesic. Mice were allowed to recover for 5 days before experimentation. Chondroitinase ABC or penicillinase (both from Sigma) was injected at 0.025 U/side 24 h before fear conditioning or 3–4 days before electrophysiological recording. In all injection experiments APP/PS1 transgenic mice and wildtype littermate controls were used at 3 months of age.

All experiments were carried out in a fear conditioning system (TSE Systems) as previously described [[32]]. Training and testing was performed in a Plexiglas chamber with a stainless steel grid floor with constant illumination (100–500 lx) and background sound (white noise, 68 dB sound pressure level), situated in a gray box to shield it from the outside. The chamber was cleaned with 70% ethanol before each session. Training consisted of placing mice in the chamber for a period of 180 s, after which a 2 s foot shock (0.7 mA) was delivered through the grid floor. Mice were returned to their home cage 30 s after the end of the shock. The retrieval tests consisted of a 180 s re-exposure to the context (conditioned stimulus) 24 h after acquisition. Baseline inactivity, exploration, distance traveled and freezing were assessed automatically. Freezing was defined as lack of any movement besides respiration and heart beat during 4 s intervals and is presented as a percentage of the total test time. Significance was determined using a Student t-test (untreated animals) or Two-Way ANOVA (penicillinase and chondroitinase ABC treated animals) followed by one-sided Student t-test post-hoc analysis.

Spatial memory was tested in a Morris water maze setup. Before testing, mice were handled for 5 days. A circular pool (ø 1.2 m) was filled with water which was painted white with non-toxic paint and kept at a temperature of 25°C. An escape platform (ø 9 cm) was placed at 30 cm from the edge of the pool submerged 0.5 cm below the water surface. Visual cues were located around the pool at a distance of ~1 m. During testing lights were dimmed and covered with white sheets and mice were video-tracked using Viewer (BiobServe, Fort Lee, NJ). Mice were trained for 5 consecutive days using 4 trials/day with a 30–180 s inter-trial interval. In each trial, mice were first placed on the platform for 30 s, and then placed in the water at a random start position and allowed a maximum of 60 s to find the platform. Mice that were unable to find the platform within 60 s were placed back on the platform by hand. After 30 s on the platform mice were tested again. To prevent hypothermia mice were placed in their home cage for 3 min between trials 2 and 3. On day 6 a probe trial was performed with the platform removed. Mice were placed in the pool opposite from the platform location and allowed to swim for 60 s. During training trials, the latency to reach the platform was measured; in the probe trial, the time spent in each quadrant of the pool was measured. Learning was assessed as the amount of time spent in the target quadrant. Significance was determined using a paired Student t-test (for learning within genotypes) or an unpaired Student t-test (for differences between genotypes).

A planar multi-electrode recording setup (MED64 system; Alpha Med Sciences, Tokyo, Japan) was used to record field excitatory post-synaptic potential (fEPSP) and to elicit LTP as previously described [[33]]. Animals were decapitated and brains were rapidly removed and placed in ice-cold slice buffer (124 mM NaCl, 3.3 mM KCl, 1.2 mM KH₂PO₄, 7 mM MgSO₄, 0.5 mM CaCl₂, 20 mM NaHCO₃ and 10 mM glucose; constantly gassed with 95% O₂/5% CO₂). Coronal hippocampal slices were prepared using a vibrating microtome at 400 μm and then placed in a chamber containing artificial cerebrospinal fluid (aCSF; 124 mM NaCl, 3.3 mM KCl, 1.2 mM KH₂PO₄, 1.3 mM MgSO₄, 2.5 mM CaCl₂, 20 mM NaHCO₃ and 10 mM glucose; constantly gassed with 95% O₂/5% CO₂). Slices were allowed to recover for 1 h and then placed on 8×8 multi-electrode arrays containing P5155 probes (Alpha Med Sciences; inter-electrode distance 150 μm) pre-coated with polyethylenimine (PEI; Sigma). After addition of 500 μl aCSF the array was placed in a moist chamber that was constantly gassed with 95% O₂/5% CO₂ for at least 1 h before recording. Correct placement of the electrodes at the CA3–CA1 region was done manually, monitored by a microscope (SZ61; Olympus, Japan). During recording, slices were constantly perfused with oxygenated aCSF containing 10 μM glycine at a flow rate of 2 mL/min at RT. fEPSPs were recorded from multiple electrodes in the stratum radiatum of CA1. An external concentric bipolar electrode (CBCBG75; FH-Company, Bowdoin, ME) in the Schaffer collateral pathway was used as the stimulating electrode using a homemade model 440b isolated Bipolar Current Stimulator. Based on the stimulus–response curve, a stimulation intensity was used that evoked fEPSPs with a magnitude of 50% of the maximum response (usually ~100 μA). After allowing a stable baseline of 10 min, LTP was evoked by a 2× 100 Hz stimulus of 1 sec each with a 15 sec interval and fEPSP responses were recorded for 1 h after the tetanus. LTP was expressed as the change in the slope of the fEPSP relative to baseline and averaged for multiple electrodes (usually 5) located in the stratum radiatum. Statistical significance was determined using a Two-Way ANOVA on the average fEPSP change from 10–20 min after LTP induction followed by Student t-test post-hoc analysis.

We first determined the progression of AD-like pathology and memory impairments in APP/PS1 mice using immunohistochemistry and behavioral analysis. In accordance with previous reports [[14],[31]], no Aβ plaques were observed at 3 months of age. At 6 months APP/PS1 mice had few hippocampal plaques, and at 12 months plaque load was further increased (Figure 1). We next evaluated hippocampal memory in 3 months old APP/PS1 mice that lack AD-like pathology. In a contextual fear-conditioning task (Figure 2a), APP/PS1 and wildtype littermates exhibited similar baseline activity during the acquisition phase before shock delivery (Figure 2b). Memory, however, was significantly reduced in APP/PS1 mice, as assessed by freezing behavior during the memory retrieval phase upon re-exposure of the animals to the context (Figure 2c). When we tested spatial memory performance in a Morris water maze task using a separate batch of animals, no differences were observed between APP/PS1 mice and wildtype littermates at 4 months of age; wildtype and transgenic animals performed equally well, both during the 5-day training (Figure 2d) and in the probe trial (Figure 2e-f). These data are in agreement with other studies showing that contextual fear memory is early affected in APP transgenic mouse models, whereas spatial and reference memory deficits only become apparent after 6–8 months [[16],[34],[35]]. We conclude that acquisition of contextual fear memory can be used as a robust and reliably measure of pre-pathological memory deficits in APP/PS1 mice.

We next asked what might be the molecular basis of early hippocampal memory impairments in APP/PS1 mice. Proteomics analysis was performed on hippocampal synaptosome fractions at 1.5 and 3 months of age, before plaque deposition, and at 6 and 12 months of age, after the onset of plaque formation (Additional file 2: Figure S1). A total of 376 proteins were quantified (Additional file 3: Table S1), of which the levels of 105 proteins were significantly different between APP/PS1 mice and wildtype controls at one or more time points (FDR <10, log-fold change >0.125). Most differentially regulated proteins, 87 in total, were detected at 3 months of age. The most differentially regulated protein detected was APP, showing significantly increased levels in APP/PS1 mice compared with wildtype controls at 3, 6 and 12 months of age, reaching a maximum of 2.5-fold increase at 12 months (Additional file 4: Figure S2a). Higher APP levels in APP/PS1 mice could to a large extent be attributed to an increase of the specific Aβ peptide LVFFAEDVGSNK, in particular at 12 months of age (Additional file 4: Figure S2b). This was confirmed by immunoblotting with an Aβ-specific antibody (Additional file 4: Figure S2c). These data demonstrate that APP/PS1 mice have an age-dependent accumulation of Aβ in hippocampal synaptosome fractions.

Typical synapse organizing proteins, such as the postsynaptic protein PSD-95 and the presynaptic proteins bassoon and piccolo, were not significantly different between APP/PS1 mice and wildtype controls, neither were the major hippocampal AMPA- and NMDA-type glutamate receptors (Additional file 3: Table S1), suggesting that other synaptic alterations underlie the observed cognitive defects in APP/PS1 mice. Functional enrichment analysis revealed the strongest enrichment at 3 months of age for proteins of the extracellular matrix (ECM; Additional file 5: Table S2). Time course expression analysis of four ECM proteins, i.e., the chondroitin sulphate proteoglycans (CSPGs) neurocan and brevican, and the proteoglycan crosslinking proteins tenascin-R and hyaluronan/proteoglycan link protein 1, revealed a gradual age-dependent upregulation in wildtype mice, which was accelerated in APP/PS1 mice resulting in higher relative expression levels at 3 and 6 months of age (Figure 3a). The initial downregulation of neurocan in wildtype mice is in line with its proposed juvenile function [[36]]. The upregulation of ECM proteins at 12 months of age in both APP/PS1 and wildtype mice marks the onset of an age-dependent increase in ECM levels that we recently also observed in normal aging mice and which continues progressively up to two years of age [[37]]. The relatively high levels of ECM proteins in APP/PS1 mice compared with wildtype controls at 3 months of age were confirmed by immunoblotting (Figure 3b). APP/PS1 mice also showed increased hippocampal staining using the ECM marker Wisteria floribunda agglutinin (WFA) [[38]] (Figure 3c-d). In particular, we observed a significant increase in the number of parvalbumin (PV)-positive cells containing WFA-positive perineuronal nets (PNNs), whereas the total number of PV-positive cells did not increase (Figure 3e).

CSPGs are the major inhibitory components of both PNNs and diffuse ECM structures [[39],[40]], and in vivo digestion of the chondroitin sulfate glycosaminoglycan side chains of CSPGs by chondroitinase ABC (ChABC) restores ocular dominance plasticity [[41],[42]], promotes recovery from spinal cord injury [[43]], and enhances erasure of fear memory [[44]], object recognition learning and long-term depression [[45]], all without adverse side effects. We therefore asked whether local injection of ChABC (0.025 U/side) into the hippocampus of 3 months old APP/PS1 mice could reverse the observed fear memory deficit. Control animals were injected with penicillinase, an enzyme with no endogenous substrate in mice. Animals were subjected to the fear conditioning task 24 h after injection (Figure 4a). APP/PS1 mice that were injected with penicillinase showed a significant reduction in fear memory compared with wildtype penicillinase treated mice (Figure 4b), and this effect was comparable to untreated mice (Figure 2c). ChABC treatment on the other hand restored memory function in APP/PS1 mice, and freezing levels were not significantly different from either ChABC or penicillinase treated wildtype mice (Figure 4b). The efficacy and regional specificity of the ChABC treatment was confirmed by post-hoc staining of hippocampal sections with WFA (Figure 4c).

At the cell physiological level contextual memory is represented as a long-term potentiation (LTP) of hippocampal synapses [[46]]. We therefore tested whether hippocampal LTP is affected in APP/PS1 mice and can be rescued by ChABC treatment. Hippocampal slices were prepared 24–48 h after injection of penicillinase or ChABC and LTP was induced by electrical stimulation of the Schaffer collateral pathway. Field excitatory postsynaptic potentials (fEPSP) were recorded in the stratum radiatum in CA1. Tetanic stimulation readily induced LTP in slices obtained from penicillinase treated wildtype mice, but LTP was significantly reduced in slices obtained from penicillinase treated APP/PS1 mice (Figure 4d-f). These findings are in accordance with previous LTP measurements in APP/PS1 mice [[47],[48]]. ChABC treatment however significantly restored LTP in APP/PS1 mice, and LTP levels were not significantly different from either ChABC or penicillinase treated wildtype mice (Figure 4f). ChABC treatment thus rescues both behavioral and physiological deficits in APP/PS1 mice at 3 months of age.