In addition to Aβ plaques and neurofibrillary tangles, Alzheimer's disease (AD) is characterized by increased brain levels of APP C-terminal fragments.
DOI 10.1007/s00702-005-0373-6 J Neural Transm (2006) 113: 1225–1241
Regional acetylcholinesterase activity and its correlation with behavioral performances in 15-month old transgenic mice expressing the human C99 fragment of APP M. Dumont1 , R. Lalonde1 , J.-F. Ghersi-Egea2 , K. Fukuchi3 , and C. Strazielle4 1
Universite de Rouen, Faculte de Medecine et de Pharmacie, INSERM U614, Rouen, and 2 Universite de Lyon-Laennec, INSERM U433, Faculte de Medecine, Lyon, France 3 Department of Biomedical and Therapeutic Sciences, University of Illinois College of Medicine, Peoria, IL, USA 4 Universite Henri Poincare, Nancy I, Laboratoire de Pathologie Moleculaire et Cellulaire des Nutriments, INSERM U724, and Service de Microscopie Electronique, Faculte de Medecine, Vandoeuvre-les-Nancy, France Received March 2, 2005; accepted August 31, 2005 Published online December 14, 2005; # Springer-Verlag 2005
Summary. In addition to Ab plaques and neurofibrillary tangles, Alzheimer’s disease (AD) is characterized by increased brain levels of APP C-terminal fragments. In the present investigation, the cholinergic innervation in forebrain regions of transgenic mice (Tg13592) expressing the human bAPP C99 fragment was compared to that of non-transgenic controls by measuring the activity of the non-specific catabolic enzyme, acetylcholinesterase (AChE). The AChE activity of Tg13592 mice was altered in several regions implicated in the functional loop of regulation between septum and hippocampus, vulnerable in Alzheimer pathology and critically involved in cognitive functions. In particular, AChE activity was upregulated in three basal forebrain regions containing cholinergic cell bodies, prelimbic cortex, anterior subiculum, and paraventricular thalamus, but downregulated in lateral septum and reticular thalamus. The increased activity in medial septum and anterior subiculum was linearly correlated
with poor performances in a spatial learning task, possibly due to cell stress mechanisms. Because of some similarities in terms of neurochemistry and behavior, this mouse model may be of use for studying prodromal AD. Keywords: Alzheimer’s disease, acetylcholinesterase, cholinergic neuron, transgenic mice, APP C99 fragment. Introduction Excitatory neuromodulation, using in part acetylcholine (ACh) with its nicotinic and muscarinic receptors is associated with learning and memory functions (Drachman and Leavitt, 1974). In cholinergic pathways, the synthetic enzyme, choline acetyltransferease (ChAT; acetyl-CoA: choline O-acetyltransferase), and the vesicular ACh transporter are localized only in the presynaptic component, whereas the non-specific catabolizing enzyme, acetylcholinesterase (AChE; acetyl-
1226
M. Dumont et al.
choline acetylhydrolase), is contained both in presynaptic cholinergic terminals and postsynaptic cholinoceptive neurons (Mesulam, 2004). Cholinergic deficits were the first neurochemical changes described in Alzheimer neocortex and hippocampus (Davies et al., 1976; Bowen et al., 1976; Perry et al., 1977; Whitehouse et al., 1982). ChAT (Coyle et al., 1983; DeKosky et al., 1992; Hardy et al., 1985) and AChE (Atack et al., 1983; Blusztajn and Berse, 2000; Coyle et al., 1983; Hardy et al., 1985; Kuhl et al., 1999; Shinotoh et al., 2000) activities decreased, while Naþ -dependent high-affinity choline uptake sites increased (Bowen et al., 1982; Bissette et al., 1996; Slotkin et al., 1994), perhaps due to compensatory mechanisms. An isoform-specific dissociation was reported for AChE, with only the tetrameric G4 and not the monomeric G1 isoform being diminished (Atack et al., 1983). The reduced activity of ChAT or AChE can be attributed in part to a loss of basal forebrain cholinergic neurons (Coyle et al., 1983; Fr€ olich, 2002; Hardy et al., 1985; Lyness et al., 2003). Postmortem studies have shown that this cell deficit is accompanied by an incremental loss of cholinergic receptors in cerebral cortical neurons, a result also observed in vitro after Ab exposure (Wevers et al., 2000). Conversely, a7 nicotinic receptors increased on astrocytes of Alzheimer cortical and hippocampic regions (Teaktong et al., 2003; Yu et al., 2005). According to the cholinergic hypothesis, the impairment of cognitive functions and the behavioral disturbances that affect patients with Alzheimer’s disease (AD) are due in part to cortical deficiencies in cholinergic neurotransmission (Baskin et al., 1999; DeKosky et al., 1992). In addition to its role in modulating synaptic acetylcholine levels, AChE can potentiate Ab aggregation and neurotoxicity (Bartolini et al., 2003; Inestrosa et al., 1996; Talesa, 2001). AChE was co-localized with neuritic and diffuse Ab plaques in brain parenchyma and in cerebrovascular walls as
well as in neurofibrillary tangles of Alzheimer brain (Gomez-Ramos et al., 1992; Kalaria et al., 1992; Tago et al., 1987; Wright et al., 1993). Co-localisation was also observed in two transgenic mouse models with the Swedish mutation of amyloid precursor protein (APP) causing Ab plaques, namely APP751SWE (TgAPP23, Boncristiano et al., 2002) and APP695SWE (Tg2576, Apelt et al., 2002). The accelerated amyloid formation in presence of AChE was independent of its catalytic activity and isoform structure (Inestrosa et al., 1996). Conversely, Ab increased AChE activity (Sberna et al., 1997), an effect that was prevented by a-tocopherol and nitric oxide synthase inhibitors, indicating mediation of this effect through oxidative stress (Melo et al., 2003). Despite extensive data, there is no consensus as to how early the cholinergic impairment occurs and to what extent it is dependent on the degeneration of the nucleus basalis of Meynert. Recent results in positron emission tomography studies indicate earlyonset cholinergic alterations (DeKosky et al., 2002; Herholtz et al., 2004). These investigations can be supplemented with transgenic mouse models reproducing preclinical and early stages of amyloid pathology. Some cholinergic deficits have been shown to appear in these models. Region-specific modifications in AChE activity were reported in APP695LD (London V642I mutation) transgenic mice with Ab plaques, being decreased in subiculum but increased in dentate gyrus and CA1 subregions of the hippocampal formation (Bronfman et al., 2000). AChE activity was unchanged in APP751SWE (Boncristiano et al., 2002) and APP695SWE (Apelt et al., 2002; Fodero et al., 2002) transgenic mice despite extensive Ab plaques. However, when specific AChE isoforms were taken into account, the activity of an abnormally glycosylated G1 version increased in cortical extracts of APP695SWE mice, whereas the activity of the tetrameric G4 version was unchanged (Fodero et al., 2002). The same
AChE activity in transgenic mice
dissociation was revealed in transgenic mice expressing the C100 fragment of human APP, where increased AChE activity was reported (Sberna et al., 1998). In order to study the potential impact of early-stage amyloid pathology on forebrain cholinergic innervation, a detailed cartography of AChE activity was undertaken in the brains of transgenic mice (Tg13592) expressing an 99-amino acid C-terminal fragment of human APP (APP=C99). Tg13592 mice presented high levels of b-amyloid peptides (Ab) in plasma, skeletal muscles, intestine, pancreas, and brain (Fukuchi et al., 1996, 1998; Strazielle et al., 2004). However, up to the age of 29 months, these mice contained Ab fibrils in all of these tissues except in brain, presumably because of the nature of the promoter (Fukuchi et al., 1998). Therefore, this model offers the opportunity of estimating brain cholinergic function caused by APP -C99 and soluble Ab peptide overexpression without b-amyloidosis. Because APP fragments and Ab peptides accumulated intracellularly in patients with hereditary AD (McPhie et al., 1997), because of their neurotoxic properties and their role on AChE activity in cell cultures (Fukuchi et al., 1993; Suh and Checler, 2002; Sberna et al., 1997), it is hypothesized that cholinergic transmission may be affected by them. The functional consequences were also analyzed by correlating enzymatic activity with the hypoactivity observed in automated chamber and elevated plus-maze tests and with impaired spatial learning in the Morris water maze previously reported in these mice (Lalonde et al., 2002). Materials and methods Animals APP=C99 (Tg13592) mice were generated and backcrossed to the C57BL=6J strain for more than nine generations at the University of Alabama (Fukuchi et al., 1996, 1998). Fifteen-month-old Tg13592 mice (n ¼ 17) and age- and sex-matched non-transgenic controls (n ¼ 19) were then shipped by air-freight to the
1227
University of Rouen (France) for behavioral analyses. This investigation conformed with the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (no 85-23, revised 1996), the recommendations edited by the European Community Council for the Ethical Treatment of Animals (no 86=609=EEC), and the regulations at each university. The animals were kept in group cages with woodchip bedding in a temperature – and humiditycontrolled room with a light-dark cycle of 12=12 hrs (lights on at 7:00) and with food and water ad libitum. Behavioral performances in motor activity and spatial tests were evaluated during a 16-day period (Lalonde et al., 2002) before histochemical assessment performed at the University of Nancy.
Tissue preparation The mice were killed by decapitation one day after behavioral testing. The brains were rapidly removed and immediately snap-frozen in isopentane at 40 C. After storage at 80 C, the tissues were serially cut into 20 mm-thick coronal sections with a cryostat HM505E (Microm Francheville, France) mounted on gelatinchrome alum-coated slides, and conserved at 80 C until processing.
AChE Histochemistry All of the chemical products used in histochemical and biochemical experiments were purchased from SigmaAldrich Chemical Company (Saint Quentin Fallavier, France). AChE histochemistry was performed on a series of dual sections from each mouse brain together with several complete sets of standards, according to the protocol originally described by Koelle and Friedenwald (1949) and presented in further detail by Paxinos and Watson (1986). Sections were incubated in Coplin jars in an incubation medium composed of 50 mM acetate buffer with 4 mM copper sulfate and 16 mM glycine, together with 3 mg ethopropazine and 116 mg S-acetylthiocholine iodide per 100 ml. The final pH value was adjusted to 5.0. Ethopropazine was included for inhibition of unspecific cholinesterases. The sections were incubated overnight (15 hours) at room temperature. After rinsing three times in distilled water, the staining was revealed by incubation in a 1% sodium sulfide solution (pH 7.5) for 10 min. The slides were then rinsed with distilled water, fixed in a 10% formalin bath for 30 min, rinsed a second time with distilled water, dehydrated with ethanol and xylene, and mounted with Eukitt. Control sections incubated without substrate always remained unstained. A similar result was obtained with the addition of 50 mM ambenonium (Tocris, Illkirch, France), a potent AChE inhibitor, into the
1228
M. Dumont et al. 25 ml of DTNB (0.5 mM) in a spectrophotometric cuvette. After recording the baseline absorbance, 10 ml of acetylthiocholine-iodide (0.5 mM) was added and the rate of absorbance increase was measured. The signal was proportional to the amount of added homogenate (r ¼ 0.997). The specific activity was 13.0 0.66 mmol=min=g of tissue (mean SD, n ¼ 8), calculated with the molar extinction coefficient of 1.36 104 Mol1 cm1 for the thionitrobenzoate anion. AChE activity remained constant in 80 C frozen standards for several months.
Densitometric analyses
Fig. 1. Acetylcholinesterase (AChE) labelling of forebrain regions (a) localized at 1.54 mm anterior to bregma plane and signal extinction after adding 50 mM ambenonium in the incubation medium (b). Note the total absence of labelling resulting from the AChE inhibitor treatment. Scale bar 0.1 mm incubation medium, confirming the specificity of AChE labelling (Fig. 1). Standards for AChE activity were prepared in cylindrical microtubes filled by mouse whole brain homogenates, frozen and kept at 80 C until cut in the cryostat at the same time as brain slices. Sections were cut at 10, 20, 30, and 40 mm in thickness. Under our experimental conditions, the intensity of staining was linearly proportional to the thickness of the standard sections (r ¼ 0.996). The specific activity of AChE was measured on a Cary 100 spectrophotometer at 22 C by a modification of the Ellman et al. (1961) colorimetric method. Thiocholine, formed during hydrolysis of acetylthiocholine, rapidly reacts with 5-50 -dithiobis-2-nitrobenzoic acid (DTNB) to release a colored 5-thio-2-nitrobenzoate anion with maximum absorption at 412 nm. 50 mg of frozen standard was homogenized in 1 ml of 0.1 M phosphate buffer (pH 8.0). Aliquots of the resulting homogenate (12.5 to 50 ml) were diluted with phosphate buffer containing
Mapping of brain AChE activity was assessed by densitometry on a BIOCOM computer-assisted image analysis system (Les Ulis, France). The optical densities were converted by means of standards into enzymatic activity in mmol=min=g of tissue. The entire mouse brain could be observed and sampled on a series of two slides, 15–18 sections per slide corresponding to sequential coronal planes at 250–350 mm. The brain regions measured were identified with the Franklin and Paxinos mouse atlas (1997). For each region, multiple optical density readings (10–50) were made by a single experienced investigator under blinded conditions with magnifications of 20 or 40 depending on the heterogeneity of the tissue and the precision needed for seeing the structures. To avoid contamination of adjacent structures and to obtain an homogeneous evaluation, measures with the same surface area were obtained for every site and animal (Reader and Strazielle, 1999). Each structure was evaluated on a single or preferably multiple sections, depending of the extent of its surface in order to obtain an average measure of labelling density. The paired structures were measured equally from each side of coronal sections of the brain and the values averaged. Thus, all of the measurements were globally performed on 19 control and 17 transgenic mice except when regions presented histological artifacts or for the small and heterogeneous structures presenting a field difference from the chosen plane. These restrictions limited the measurements to no less than 14 to 15 mice per group.
Brain pathology detection Sections adjacent to those labelled for AChE histochemistry were stained with a 0.5% cresyl violet solution for precise delimitation of subregions and for assessment of possible brain atrophy in transgenic mice. Neuronal degeneration in forebrain cholinergic regions was detected on a new series of sections from each brain by using Fluoro-Jade B, according to the protocol described by Schmued and Hopkins (2000). The slides were evaluated under fluorescent (FITC) microscopy
AChE activity in transgenic mice (Olympus AX70, France) for the detection of fluorescent yellow-green cells, labelling dying neurons.
Statistical analyses To take into account a high number of comparisons, a two-way repeated ANOVA was used to compare mean group differences at an a level of P ¼ 0.05: comparisons were performed among the constitutive subregions of each structure or among regions of a functional brain circuitry. The unpaired-t test, with a criterion set at P ¼ 0.01, was then used for individual group comparisons in those brain regions with significant ANOVAs. Linear correlations were undertaken in both groups of mice for AChE activity measured between different subregions of cerebral cortex, neostriatum, basal forebrain cholinergic nuclear complex, thalamus, and hippocampus, as well as interregionally, between the above-named structures and cholinergic nuclei. For the transgenic group, AChE was also correlated with behavioral perfomances (Lalonde et al., 2002), but only for brain regions showing intergroup differences. For correlational analyses, the threshold of significance was set at P ¼ 0.05.
Results AChE enzymatic activity The distribution of AChE activity was heterogenous (Table 1), with elevated activity noted in Tg13592 mice particularly but not restricted to the forebrain cholinergic system, including specific limbic regions interconnected with some forebrain structures. Cerebral cortex The AChE staining pattern in mouse neocortex was diffuse and weak in the neuropil, with moderate staining bands in the molecular layer, occupying the deeper part of layer IV and=or the upper part of V depending on the cortical area. Among cortical subregions, only the prefrontal area showed a transgene effect (F1,23 ¼ 4.3; p