M1 muscarinic acetylcholine receptor dysfunction in moderate Alzheimer’s disease pathology

Abstract Aggregation of amyloid beta and loss of cholinergic innervation in the brain are predominant components of Alzheimer’s disease pathology and likely underlie cognitive impairment. Acetylcholinesterase inhibitors are one of the few treatment options for Alzheimer’s disease, where levels of available acetylcholine are enhanced to counteract the cholinergic loss. However, these inhibitors show limited clinical efficacy. One potential explanation for this is a concomitant dysregulation of cholinergic receptors themselves as a consequence of the amyloid beta pathology. We tested this hypothesis by examining levels of M1 muscarinic acetylcholine receptors in the temporal cortex from seven Alzheimer’s disease and seven non-disease age-matched control brain tissue samples (control: 85 ± 2.63 years old, moderate Alzheimer’s disease: 84 ± 2.32 years old, P-value = 0.721; eight female and six male patients). The samples were categorized into two groups: ‘control’ (Consortium to Establish a Registry for Alzheimer’s Disease diagnosis of ‘No Alzheimer’s disease’, and Braak staging pathology of I–II) and ‘moderate Alzheimer’s disease’ (Consortium to Establish a Registry for Alzheimer’s Disease diagnosis of ‘possible/probable Alzheimer’s disease’, and Braak staging pathology of IV). We find that in comparison to age-matched controls, there is a loss of M1 muscarinic acetylcholine receptors in moderate Alzheimer’s disease tissue (control: 2.17 ± 0.27 arbitrary units, n = 7, Mod-AD: 0.83 ± 0.16 arbitrary units, n = 7, two-tailed t-test, t = 4.248, P = 0.00113). Using a functional rat cortical brain slice model, we find that postsynaptic muscarinic acetylcholine receptor function is dysregulated by aberrant amyloid beta-mediated activation of metabotropic glutamate receptor 5. Crucially, blocking metabotropic glutamate receptor 5 restores muscarinic acetylcholine receptor function and object recognition memory in 5XFAD transgenic mice. This indicates that the amyloid beta-mediated activation of metabotropic glutamate receptor 5 negatively regulates muscarinic acetylcholine receptor and illustrates the importance of muscarinic acetylcholine receptors as a potential disease-modifying target in the moderate pathological stages of Alzheimer’s disease.


Introduction
Alzheimer's disease causes neurodegeneration in the brain and represents the major, and as yet incurable, form of dementia. A striking feature of the disease is the loss of cholinergic innervation of the cortex, which occurs alongside the pathological progression of Alzheimer's disease. The 'cholinergic hypothesis of Alzheimer's disease' has accordingly directed the development of clinically approved acetylcholinesterase inhibitors that function to enhance available acetylcholine in synapses. However, acetylcholinesterase inhibitors are fundamentally limited in their clinical efficacy (Bond et al., 2012), indicating that the loss of acetylcholine alone does not account for the cholinergic dysfunction. Thus, whilst acetylcholine abundance is doubtlessly important, functional impairment of the acetylcholine receptor itself in the synapse may also play a role in Alzheimer's disease. Therefore, it is of interest to determine whether muscarinic acetylcholine receptor (mAChR) dysfunction features in Alzheimer's disease, and if so whether this receptor can be targeted as a diseasemodifying strategy.
The M1 mAChR is a sub-family of G-protein-coupled mAChRs. They operate through a Gi/Gq signalling cascade, mobilizing Ca 2þ from internal Ca 2þ stores (Caulfield, 1993). The activation of mAChRs induces long-term synaptic plasticity in the perirhinal cortex, an area pivotal for novel object recognition memory (Warburton et al., 2003). Indeed, blocking mAChRs causes deficits in object recognition memory (Warburton et al., 2003). Given that this particular form of memory is highly sensitive to disruption in the early stages of Alzheimer's disease (Dickerson et al., 2017), and the perirhinal cortex is subject to volume loss in Alzheimer's disease (Juottonen et al., 1998), this provides an intriguing putative link between impaired mAChR function and memory deficits in Alzheimer's disease.
At the molecular level, activation of mAChRs negatively regulates synaptic N-methyl-D-aspartate receptor (NMDAR) function (Jo et al., 2010) and glycogen synthase kinase 3 signalling (De Sarno et al., 2006), a key component in a mechanism that drives tauopathy in the brain (Hernandez et al., 2012). Critically, M1 mAChR function is inhibited by aberrant activation of metabotropic glutamate receptor 5 (mGluR5) in the perirhinal cortex (Jo et al., 2006). This is important as mGluR5 has been shown to act as a co-receptor for amyloid beta (Ab) at the postsynaptic terminal (Um et al., 2013;Haas and Graphical Abstract Strittmatter, 2016), playing an intermediary role between Ab and intracellular signalling cascades that lead to synapse loss and neurodegeneration (Um et al., 2013;Haas and Strittmatter, 2016). Therefore, mGluR5 may be involved in the pathological impairment of mAChR in the perirhinal cortex in Alzheimer's disease.
To test this hypothesis, we first investigated whether M1 mAChR, the dominant form of mAChR in the brain, is aberrantly regulated in the moderate pathological stages of Alzheimer's disease. We find a reduction in M1 mAChR protein expression in post-mortem temporal cortical tissue from patients with moderate-stage Alzheimer's disease. We also find that Ab oligomers facilitate the aberrant activation of mGluR5, consequently leading to weakening of postsynaptic M1 mAChR function in the perirhinal cortex. Furthermore, systemic infusion of an mGluR5-negative allosteric modulator [LSN2300979 (LSN)] rescues the otherwise impaired object recognition memory in the 5XFAD (familial Alzheimer's disease) transgenic (TG) mouse model of Alzheimer's disease pathology. Thus, our results suggest that inhibition of mGluR5 is beneficial for mAChR function and disease modification in moderate stages of Alzheimer's disease.

Materials and methods
Post-mortem brain tissue Human post-mortem brain tissue was provided by the South West Dementia Brain Bank (University of Bristol, UK), with local Research Ethics Committee approval. A total of 200 mg of non-specified temporal cortex samples from 14 patient brain tissue donations were used for this study (8 female and 6 male patients). The samples were categorized as 'control' or 'moderate Alzheimer's disease'. All control samples had a CERAD diagnosis (Mirra et al., 1991) of 'No Alzheimer's disease', andBraak staging (Braak andBraak, 1995) pathology of I-II. All moderate Alzheimer's disease samples had a CERAD diagnosis of 'possible/probable Alzheimer's disease', and Braak staging pathology of IV. No differences in age (control: 85 6 2.63 years old, moderate Alzheimer's disease: 84 6 2.32 years old, P-value ¼ 0.721, 'data not shown') or 'post-mortem' delay (control: 45.78 6 4.78 h old, moderate Alzheimer's disease: 44.28 6 11.34 h, Pvalue ¼ 0.209, 'data not shown') were detected between control and moderate Alzheimer's disease groups. Human 'post-mortem' frozen samples were lysed at room temperature in sucrose/sodium dodecyl sulfate (SDS) buffer (lysis buffer) containing 1% SDS, 0.1 g/ml sucrose, 1 mM Ethylenediaminetetraacetic acid (EDTA), protease inhibitor cocktail at 1:10 (05892791001, Roche), phosphatase inhibitor cocktail at 1:100 (P5726, Sigma) and the serine/ cysteine protease inhibitor phenylmethanesulfonyl fluoride at 1:100 (PMSF, P7626; Sigma). Approximately 100 mg of tissue were lysed in 500 ll of sucrose/SDS buffer. Samples were then centrifuged at 16 g for 20 min at 16 C. Pellets were discarded, and supernatants were collected in a fresh tube, constituting the stock solutions. Stock solutions were diluted 1:10 in lysis buffer to generate the working concentration sample. Samples were aliquoted and stored at À80 C until required. Protein levels were quantified using bicinchoninic acid (BCA) assay (Thermo Fisher Scientific). Following electrophoresis using sodium dodecyl sulphate-polyacrylamide gels, immunoblot analysis were performed using the relevant antibodies and visualized with chemiluminescence. The chemiluminescent signal was detected using the G:Box Chemi XT4 imaging system (Syngene, UK), and images were taken with GeneSys software (Syngene, UK). The following primary antibodies and dilutions were used: Anti-NR1 (1:1000, Upstate, 06-31; Matsuno et al., 2015), Anti-GluA2

Animals
All procedures involving animals used for in vitro experiments were performed in accordance with the UK Animals (Scientific Procedures) Act, 1986. All experimental protocols were approved by the University of Bristol Animal Welfare and Ethical Review Body. Animals were housed in groups and exposed to the 12-h light/12-h dark cycle. Animals were housed in controlled environmental conditions with food and water available ad libitum. Calcium imaging experiments were conducted using organotypic slices prepared from P7-8 male Wistar rats. Electrophysiology experiments were conducted using acute slices prepared from P25-35 male Wistar rats. Behavioural experiments were approved by the Institutional Animal Care and Use Committee of Kyung Hee University (Seoul, Republic of Korea). Mice were housed four per cage and were allowed access to water and food ad libitum. Animals were kept at a constant environment; temperature (23 6 1 C) and relative humidity (60 6 10%) under a 12-h light/dark cycle (lights on from 7:30 A.M. to 7:30 P.M.). Behavioural experiments were conducted using 6-month-old 5XFAD TG mice. These animals overexpress mutant amyloid precursor protein [K670N/M671L (Swedish), I716V (Florida) and V717I (London mutation)], along with mutant PS1 with M146L and L286V mutations. This results in the enhanced production of Ab with a rapid onset, with Ab typically expressed intraneuronally at 1-2 months of age (Oakley et al., 2006).

Electrophysiology
Whole cell recording for holding current was performed using acute perirhinal cortex slices (400 lm), which were perfused with artificial CSF (continuously bubbled with 95% O 2 /5% CO 2 ) containing 124 mM NaCl, 3 mM KCl, 26 mM NaHCO 3 , 1.25 mM NaH 2 PO 4 , 2 mM CaCl 2 , 1 mM MgSO 4 and 10 mM D-glucose. Recording electrodes (5-6 MX) containing potassium gluconate filling solution were used to patch and voltage clamp rhinal sulcus, layer II/III neurons. These were visualized using oblique contrast microscopy, through an Olympus BX51-WU microscope (Olympus, UK). The holding current recording experiments were perfused with MK-801 (10 lM) and tetrodotoxin (500 nM) in artificial CSF. The baseline of holding current was recorded for 10 min and then carbachol (CCh; 50 lM) applied for 5 min in the perfusate.
The 1,1,1,3,3,3-hexafluoro-2-propanol/peptide solution was next dried under a gentle stream of nitrogen gas. Then 100% dimethyl sulfoxide (DMSO) was then used to resuspend the peptide, which was then incubated at room temperature for 12 min with occasional vortexing. The final solution was aliquoted into smaller volumes and stored at À80 C. For a working solution, 500-1000 ml (depending on the final concentration to be used) Dulbecco's phosphate-buffered saline (D-PBS; Invitrogen, UK) was added to the peptide stock solution and incubated for 2 h at room temperature to allow for peptide aggregation. We have previously characterized this preparation to contain 1-5 nM of low-n oligomers (Whitcomb et al., 2015).

Organotypic brain slice culture
Perirhinal cortex organotypic slice cultures were prepared from 7-to 8-day-old male Wistar rats. After decapitation, the brain was immediately placed in ice-cold cutting solution (in mM: 238 sucrose, 2.5 KCl, 26 NaHCO 3 , 1 NaH 2 PO 4 , 5 MgCl 2 , 11 D-glucose and 1 CaCl 2 ). A midsagittal section of the brain was made, the rostral and caudal parts were removed by single scalpel cuts 45 to the dorsoventral axis and each half was glued by its caudal end to a vibroslice stage (Campden Instruments, Sileby, UK). Slices (350 lm) that included perirhinal, entorhinal and temporal cortices were trimmed to create sections centred on the rhinal sulcus, containing layers I-V. Following washing, slices were plated on the top of semipermeable membrane inserts (Millipore Corporation, Bedford, MA, USA) within a six-well plate, which contained culture medium, comprised 78.8% minimum essential medium, 20% heat-inactivated horse serum, 25 mM HEPES, 10 mM D-glucose, 26 mM NaHCO 3 , 2mM CaCl 2 , 2 mM MgSO 4 , 70 mM ascorbic acid, 1 mg/ml insulin, pH adjusted to 7.3 and 320-330 mOsm. The slices were then cultured in an incubator (35 C, 5% CO 2 ) for 7-10 days in vitro. The medium was changed every 2 days. Neurons were transfected with GCaMP6 using a biolistic gene gun (Helios Gene-gun system; Bio Rad, USA) at days in vitro 3-4. Imaging assays were performed 3-4 days after transfection.

Calcium imaging
Cultured perirhinal cortex slices were biolistically transfected with GCaMP6, the genetically encoded calcium indicator. Fluorescent images were captured using a BX51-WU microscope fitted with an ORCA-ER digital camera (Hamamatsu), at 1 frame/5 s. Following a 2 min baseline, 10 ml, 50 lM CCh was perfused onto the slice using a gravity-fed dropper system. Fluorescence intensity was quantified as a percentage of the normalized baseline, using Image J (National Institutes of Health, Bethesda, MD, USA).

Object recognition
The experimental apparatus consisted of a black rectangular open field (25 cm Â 25 cm Â 25 cm). The novel object recognition task was carried out as described elsewhere (Li et al., 2004;Barker and Warburton, 2011) with slight modifications. For the novel object location recognition test, mice were habituated to the open field in the absence of objects but with an internal cue on one of the four walls for 4 days. In the training phase, mice were replaced in the same box but now with two distinct objects. The objects consisted of a glass box and plastic cylinder. Mice were allowed to freely explore the environment and the objects for 5 min. After 3 h, mice were placed back in the rectangular environment for the testing phase. The two objects were again present, but one of the two objects were now displaced to a novel spatial location. Mice were allowed to freely explore the environment and the objects for 5 min. Time spent exploring the displaced and non-displaced objects were measured. Each group's ability to recognize the novel location was determined by dividing the mean time exploring the novel object location by the mean of the total time exploring both objects during the test session. This value was multiplied by 100 to obtain a percentage preference for the novel object (T novel /T total Â 100, T familiar /T total Â 100). 5XFAD TG mice were treated with oral administration of LSN (30 mg/kg) or vehicle (Veh) (0.9% saline) for 8 days. On Days 4 and 5, animals from both groups were habituated to test environment. Following training on Day 6, animals were tested on Day 7, or in the 30day interval trial training took place on Day 37 and testing took place on Day 38.

Statistical analysis
No statistical methods were used to predetermine sample sizes. However, sample sizes were chosen based on previous studies published in the field by this laboratory and others. No data points were excluded from the analysis. Data pooled across slices are expressed as the mean 6 SEM. Statistics were based on two-tailed Student's t tests or one-way ANOVA followed by Tukey's post hoc analyses for multiple comparisons, unless otherwise indicated. P < 0.05 was considered statistically significant.

Data availability
The data pertaining to the findings described here are available from the corresponding author upon reasonable request.

Ab impairs mAChR function
Organotypic cultured perirhinal cortex slices were biolistically transfected with the genetically encoded Ca 2þ indicator GCaMP6 (Chen et al., 2013), and we measured the magnitude of fluorescence induced by Ca 2þ mobilization on ligand activation of mAChRs. Bath application of CCh (50 lM), the mAChR agonist, induced a robust and transient increase in Ca 2þ both in the soma (open symbols, peak fluorescence intensity as a % of baseline, 205 6 16%, n ¼ 13 cells from 10 slices) and dendrites (closed symbols, peak fluorescence intensity as a % of baseline 266 6 23%, n ¼ 8 dendrites from eight slices; Fig. 2A). However, in slices treated for 2 h with 1-5 nM Ab oligomers, the CCh-induced Ca 2þ signal was significantly reduced (open symbols, soma peak fluorescence intensity as a % of baseline 145 6 10%, n ¼ 16 cells from 12 slices; closed symbols, dendrite peak fluorescence intensity as a % of baseline 143 6 12%, n ¼ 8 dendrites from 9 slices; Fig. 2B). The magnitude of CCh-induced fluorescence was reduced in Ab-treated slices compared with control slices, in both the soma (control versus Ab, two-tailed t-test, t ¼ 3.320, P ¼ 0.00258) and dendrites (control versus Ab, two-tailed t-test, t ¼ 4.847, P ¼ 0.000213; Fig. 2C). This suggests that Ab oligomers rapidly impair mAChR function.

Weakening of postsynaptic mAChR function by Ab-induced pathophysiology
To confirm this, we next analysed postsynaptic mAChRmediated holding current. Neurons were voltage clamped at À70 mV, and the injected current required to maintain the holding current was monitored, before and after application of CCh (50 lM). Perfusion of CCh caused a rapid increase in negative holding current, indicative of activation of mAChRs and consequent inward current due to Ca 2þ release from intracellular stores (peak depression À62 6 14 pA, n ¼ 5; Fig. 2D). In slices treated for 2 h with 1-5 nM Ab oligomers, there was a significant reduction in holding current change in response to CCh (peak depression À24 6 3 pA, n ¼ 5; Fig. 2E; control versus Ab P ¼ 0.0368), indicating that postsynaptic mAChRs were no longer activated to the same extent as that observed in the control condition. This suggests that Ab dysregulates postsynaptic mAChR function, possibly through the impairment of classical G-protein-coupled receptor-mediated signal cascades or through the modulation of the expression of the receptor itself. Since the aberrant mobilization of Ca 2þ by Ab is known to trigger neurotoxicity in cell models (Demuro et al., 2005), we wondered whether extracellular Ca 2þ was important in this effect. To test this, slices were exposed for 2 h with 1-5 nM Ab oligomers in Ca 2þ -free artificial CSF and CCh-induced holding current change was examined. Interestingly, we found that under these conditions, Ab had no impairing effects (À63.8 6 14.3 pA, n ¼ 10; Fig. 2F). This suggests that Ab-mediated Ca 2þ mobilization causes mAChR dysfunction. The NMDAR is a central postsynaptic Ca 2þ regulator and a key conduit for Ab-induced neuronal dysfunction (Wang and Reddy, 2017). Furthermore, the inhibition of NMDAR is beneficial in Alzheimer's disease and object recognition memory in animal models of the disease (Romberg et al., 2012). Therefore, we hypothesized that Ab-induced impairment of mAChR function occurs in an NMDAR-dependent manner. When slices were co-exposed to 1-5 nM Ab oligomers and MK-801 (10 lM), the NMDAR channel blocker, in normal Ca 2þ -containing artificial CSF for 2 h, we found that Ab did not impair the inward-holding current change (À77.2 6 12.7 pA, n ¼ 7; Fig. 2G). Together, these data suggest that both extracellular Ca 2þ and activation of NMDARs are essential for Ab-mediated mAChR dysfunction.

Ab inhibits postsynaptic mAChR function by aberrant activation of mGluR5
The function and expression of mAChRs is regulated by a number of different cellular processes, including phosphorylation and trafficking (van Koppen and Kaiser, 2003). A previous study also found that juvenile rat mAChR function was down-regulated by aberrant activation of mGluR5, a group I metabotropic glutamate receptor (Jo et al., 2006). Furthermore, evidence suggests that mGluR5 and the synergistically regulated NMDAR function may be involved in Ab-mediated pathology and deficits to object recognition memory in Alzheimer's disease (Alagarsamy et al., 2001;Renner et al., 2010;Hu et al., 2014;Hamilton et al., 2016). We therefore hypothesized that the weakening of mAChR function in Alzheimer's disease pathology is underpinned by the aberrant interplay between NMDAR and mGluR5 activation. We applied the mGluR5 antagonist 2-methyl-6-(phenylethynyl)pyridine (MPEP, 50 lM) to slices and then exposed those slices for 2 h to 1-5 nM Ab oligomers. We found that, in contrast to Ab treatment alone (Fig. 2B), inhibition of mGluR5 restored CCh-induced Ca 2þ signalling (triangle symbols, soma peak fluorescence intensity as a % of baseline 254 6 37%, n ¼ 7 cells from five slices/four animals; dendrite peak fluorescence intensity as a % of baseline 310 6 49%, n ¼ 5 dendrites from 5 cells; Fig. 3A). The magnitudes of somatic and dendritic Ca 2þ on CCh application did not differ between the Ab with MPEP group and untreated controls (soma: control versus Ab þ MPEP, two-tailed t-test, t ¼ À1.432, P ¼ 0.169; dendrite: control versus Ab þ MPEP, twotailed t-test, t ¼ À0.903, P ¼ 0.386; Fig. 3B). Consistent with these findings, we found that the application of MPEP or the mGluR5-negative allosteric modulator LSN (10 lM) prevented the Ab-mediated impairment of CChinduced inward-holding current change (peak depression À79 6 7 pA, n ¼ 6, control versus Ab þ MPEP P ¼ 0.3; Fig. 3C; peak depression À71.5 6 14.5 pA, n ¼ 6; Ab þ LSN; Fig. 3D). Together, these data indicate that the impairment to mAChR function that is induced by Ab occurs in unison with the aberrant activation of mGluR5, and this is causally linked.
An mGluR5-negative allosteric modulator restores novel object recognition memory in 5XFAD TG mouse Since the inhibition of mGluR5 was critical to maintain mAChR function, and the function of mAChRs is important in object recognition memory (Warburton et al., 2003), we wondered what role mGluR5 might play in the widely reported dysregulation of cognitive function by Ab (McGowan et al., 2006). We therefore examined novel object recognition memory in the 5XFAD TG mouse model. Whilst there were no differences in exploration time between experimental groups [one-way ANOVA (Tukey's post hoc test) P ¼ 0.1789, wild type (WT): 20.23 6 4.665 s, n ¼ 8; TG-Veh: 7.946 6 2.279 s, n ¼ 7; TG-LSN: 12.01 6 6.068 s, n ¼ 7, Fig. 3E], there was a significant impairment in object discrimination performance in the  (N ¼ 7). Dots represent the value for each subject, and error bars represent the SEM. Significance level (*P < 0.05) was determined using an unpaired two-tailed Student's t-test. Insets show cropped representative immunoblots, full length blots available in Supplementary Fig. 1  transfected neurons from perirhinal cortex organotypic culture. Following a 1-min baseline, CCh is perfused onto slices. Whilst this causes a robust increase in fluorescence in control slices, (B) a much smaller magnitude of fluorescence increase is found in slices treated with Ab. (C) Quantification in peak fluorescence change induces by CCh perfusion. Ab-Treated slices had statistically smaller peak fluorescence values in both soma and dendrites compared with control slices. Bars represent 6 SEM. (D) Change in holding current induced by CCh perfusion, normalized to a 10-min baseline recording. Whilst CCh induces robust change in holding current, (E) CCh induces a modest change in holding current in slices pre-exposed to Ab prior to recording (F). This impairment of holding current change by Ab is ameliorated in slices pre-exposed to Ab in Ca 2þ -free artificial CSF (G) and in slices pre-exposed to Ab in the presence of the NMDAR blocker MK-801. Figures represent pooled data across separate experiments, with error bars representing 6 SEM. Significance level (*P < 0.05) was determined by using an unpaired Student's ttest. Scale bar indicates 10 lm. (A) Quantification of GCaMP6-mediated Ca 2þ fluorescence in transfected neurons from perirhinal cortex organotypic culture. Whilst there was mAChR impairment in moderate Alzheimer's disease 5XFAD mice compared with WT mice [0.2971 6 0.07748 (n ¼ 7) versus TG-Veh: À0.2122 6 0.1565 (n ¼ 6), P < 0.05, Fig. 3E]. Critically, this deficit was ameliorated in 5XFAD mice treated with LSN [TG-Veh versus TG-LSN: 0.3889 6 0.1234 (n ¼ 6), P < 0.001, Fig. 3E]. Indeed, whilst the WT animals showed an exploration preference for the novel object (novel: 64.860 6 3.874% versus familiar: 35.140 6 3.874% object preference, Fig. 3F), the 5XFAD mice displayed no such preference (novel: 39.390 6 7.827% versus familiar: 60.610 6 7.827% object preference, Fig 3F). Importantly, the 5XFAD animals treated with LSN displayed similar object preferences to the WT control animals (novel: 69.450 6 6.169% object preference versus familiar: 30.550 6 6.169% object preference, Fig. 3F).
We wondered whether the restorative effects of mGluR5 antagonism reflected a fundamental amelioration of the disease state. To test this, we administered the same drug regime but introduced a 30-day interval period between the last drug infusion and training and testing. Whilst there were no differences in exploration time between experimental groups [one-way ANOVA (Tukey's post hoc test) P ¼ 0.275, WT: 18.69 6 3.975 s, n ¼ 8; TG-Veh: 11.39 6 3.994 s, n ¼ 7; TG-LSN: 11.70 6 2.472 s, n ¼ 7, Fig. 3G], there was a significant impairment in object discrimination performance across the groups [oneway ANOVA (Tukey's post hoc test) P ¼ 0.0102, WT: 0.2716 6 0.09399, n ¼ 8; TG-Veh: À0.06025 6 0.05867, n ¼ 5; TG-LSN:-0.02718 6 0.03122, n ¼ 6]. However, there was no difference in discrimination performance between the 5XFAD Veh mice and those treated with LSN (P ¼ 6.134). Whilst the WT animals showed an exploration preference for the novel object (novel: 63.580 6 4.699% object preference versus familiar: 36.420 6 4.699% object preference, Fig. 3H), the 5XFAD mice displayed no such preference (novel: 46.990 6 2.933% object preference versus familiar: 53.010 6 2.933% object preference, Fig 3H). Importantly, the 5XFAD animals treated with LSN also showed no preference (novel: 48.640 6 1.561% object preference versus familiar: 51.360 6 1.561% object preference, Fig. 3H). This indicates that whilst downregulation of mGluR5 function is a putative means by which cognitive function can be restored, the pathology is underpinned by the constitutive aberrant activation of the receptor and could therefore require sustained inhibition for restorative effect.

Discussion
Our protein assay data from moderate Alzheimer's disease post-mortem brain tissue indicates a reduction in M1 mAChR. This is consistent with a dysfunction of the cholinergic transmission system occurring in the mild cognitive impairment phase of the moderate pathological stages of Alzheimer's disease (Haense et al., 2012). On the surface, our findings run counter to previous reports describing no change (Mash et al., 1985) and an increase (Overk et al., 2010) in M1 mAChRs in Alzheimer's disease human tissue. Importantly, these studies assayed tissue from patients with late-stage Alzheimer's disease, reinforcing our conclusion that mAChR dysregulation is a feature of the progressive pathology of Alzheimer's disease, and our present findings reflect moderate-stage effects. Our postnatal and juvenile rat model studies are indicative of the possible early events of Ab-mediated dysfunction of mAChR. With this in mind, our results raise the question as to whether the reduction in M1 mAChR is a consequence of pathophysiology and underlies cognitive impairment. mAChR-mediated synaptic transmission plays a key role in the cellular mechanism of object recognition in the rat perirhinal cortex (Warburton et al., 2003). Interestingly, we previously found that mGluR5 has a dominant role over M1 mAChR in this brain region in rats (Jo et al., 2006). Since M1 mAChR and mGluR5 share the same Gi/Gqcoupled signalling mechanisms, the aberrant activation of mGluR5 function may dominantly sequester G proteins, resulting in a lack of available G-protein coupling with M1 mAChR. An alternative intriguing explanation is that Ab disrupts M1 mAChR regulatory machinery, breaking down the stabilization and recycling processes of mAChR to the synapse (Gainetdinov et al., 2004). Thus, it is Figure 3 Continued only a modest increase in fluorescence in Ab-treated slices, there was a robust increase in fluorescence in Ab þ MPEP-treated slices. Scale bar indicates 10 lm. (B) There was no statistically significant difference in peak fluorescence intensity induced by CCh perfusion between control and Ab þ MPEP-treated slices. (C) Change in holding current induced by CCh perfusion. A robust change in holding current was induced by CCh in slices treated with Ab þ MPEP (D) and in slices treated with Ab þ LSN. Holding current data illustrated in C and D were recorded from acute perirhinal cortex slices from p25-35 Wistar rats. (E) Quantification of behavioural assessment of object recognition memory in 5XFAD mice, comparing WT, vehicle (5XFAD þ vehicle) and LSN-treated (5XFAD þ LSN) animals. (F) Quantification of object preference between novel and familiar objects, comparing the animal treatment groups. Whilst WT animals showed a clear preference for the novel object, 5XFAD þ vehicle animals did not. 5XFAD þ LSN animals showed a similar novel object preference to WT animals. (G) Quantification of behavioural assessment of object recognition memory across the experimental conditions outlined above, with a delay of 30 days introduced between vehicle or LSN treatment and the training and testing. (H) Whilst WT animals displayed a preference for novel objects, neither the vehicle-treated nor LSN-treated animals did. Significance level (*P < 0.05; **P < 0.01; ***P < 0.001) was determined by using a one-way ANOVA with post hoc Tukey analysis. Dots in E and G represent the value for each subject.
plausible that Ab may cause weakening of M1 mAChR function through mGluR5 activity and destabilization of mAChR. However, it is not known whether and/or how aberrant activation of mGluR5 function affects the expression of M1 mAChR in the progression of Alzheimer's disease.
Considering the slow progress in translating basic characterization of Alzheimer's disease pathophysiology to effective symptom and disease-modifying therapies, new targets are required. Synapses serve as sites where key pathological changes occur in Alzheimer's disease (Selkoe, 2002). Defects in synaptic transmission and dendritic spine structure are well-associated with the progression of pathological events that greatly contribute to the destabilization of synaptic contacts and therefore are likely to have a significant impact on memory circuits (Cochran et al., 2014). More importantly, these changes occur at early stages of the disease, a period now considered to be critical for successful therapeutic interventions (Korolev, 2014). Within this context, our key finding is that Ab drives the weakening of postsynaptic M1 mAChR function by aberrantly activating mGluR5. Since mAChR functions to affect a number of signal cascades, weakening of postsynaptic mAChR may have fundamental pathophysiological consequences, likely impacting on synaptic function and structure. For instance, weakening of mAChR may lead to aberrant activation of NMDAR in the perirhinal cortex (Jo et al., 2010). Therefore, Abmediated weakening of mAChR could also link to the regulation of a glycogen synthase kinase 3 signalling, one of the major Alzheimer's disease-related synapse-weakening molecules (Jo et al., 2011;Regan et al., 2015).
Within the limited therapeutic options available for Alzheimer's disease, memantine is a widely used NMDAR blocker. We previously reported that memantine restored object recognition memory by blocking false memory in the TG CRND8 mouse model of Alzheimer's disease (Romberg et al., 2012). This showed that the aberrant activation of NMDARs causes pathophysiological synaptic plasticity and may dysregulate cellular mechanisms of object recognition memory in Alzheimer's disease. Interestingly, our current study showed that Abmediated weakening of mAChR function was Ca 2þ mobilization and NMDAR dependent. This suggests the notion that Ab inhibits mAChR function via either mGluR5 or through synergistic regulation between the mGluR5 and NMDARs. Since mGluR5 antagonism alone rescues mAChR function and object recognition memory, tight control of mGluR5 function will be an advantageous for the maintenance of mAChR function and preventing the progress of pathology and enhancing cognitive performance, as we have shown in this study.
Though previous studies have demonstrated that a broad conceptual role of mGluR5 antagonism is the restoration of memory function in TG mouse models of Alzheimer's disease (e.g. Hamilton et al., 2016), our study provides a framework for these effects: the aberrant activation of mGluR5 impairing postsynaptic mAChR function, where through blocking mGluR5, postsynaptic mAChR function and object recognition memory in Alzheimer's disease models can be restored. Thus, we postulate that mGluR5-mediated dysfunction of postsynaptic mAChRs underlies a potential cellular mechanism of object recognition memory deficit in Alzheimer's disease. This positions mAChR as a further potential therapeutic target when considering how to prevent acetylcholine-signalling decline in the developing pathology of Alzheimer's disease.

Supplementary material
Supplementary material is available at Brain Communications online.