Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Expression in brain of amyloid precursor protein mutated in the alpha‐secretase site causes disturbed behavior, neuronal degeneration and premature death in transgenic mice.

Expression in brain of amyloid precursor protein mutated in the alpha‐secretase site causes... The EMBO Journal vol.15 no.6 pp.1265-1274, 1996 Expression in brain of amyloid precursor protein mutated in the a-secretase site causes disturbed behavior, neuronal degeneration and premature death in transgenic mice after proteolytic cleavage (reviewed in extracellular space Dieder Moechars, Kristin Lorent, Selkoe, 1994). Bart De Strooper, Ilse Dewachter and mediated by an uncharacterized This proteolysis step, Fred Van Leuven' is an intrinsic and essential proteinase named a-secretase, Experimental Genetics Group, Center for Human Genetics, secretion of APP and has become step in the mechanism of Campus Gasthuisberg, B-3000 Leuven, Belgium K.U.Leuven, of APP, both the focal point in the analysis of processing in the pathology 'Corresponding author fundamentally and for its involvement in the of Alzheimer's disease (AD). The senile plaques hallmark, A double mutation in the a-secretase site in the brain of AD patients that are a pathological PA4 region of mouse amyloid precursor protein (APP) contain mainly the ,A4 peptide (Glenner and Wong, 1984; amino acid peptide its secretion from COS cells, polarized MDCK Masters et al., 1985). This 39-43 reduced neurons. Expression of this represents part of the transmembrane and ectodomain of cells and rat primary using the neuron-specific APP (Goldgaber et al., 1987; Kang et al., 1987; Tanzi mutant in the brain of mice, et al., 1987). Since most patients with trisomy 21 develop elements of the mouse Thy-i gene promoter, resulted in AD-like senile plaques in their brain (Glenner and Wong, transgenic mice that became progressively hyperactive, 1984), this is to be attributed to a gene dosage effect of displayed seizures and died prematurely. In three the APP gene (Rumble et al., 1989). This indication for different transgenic lines the severity of the phenotype the narrow margin of regulation of expression of APP is was related directly to the expression levels of the corroborated further by the rare, early onset and presenile estimated by both mRNA and protein levels. transgene, familial AD cases that are inherited in an autosomal In addition, homozygous mice derived from each trans- dominant manner (Chartier-Harlin et al., 1991; Goate strain showed more severe symptoms which also genic et 1991; Murrell et al., 1991; Mullan et al., 1993). in life than in heterozygotes. The al., occurred earlier membrane-bound APP is cleaved not essentially dif- Newly synthesized observed symptoms were, however, intracellularly within the ,A4 peptide sequence by in ferent the different lines. Increased aggressiveness, the soluble 110-130 kDa acid and N-methyl-D- a-secretase, thereby generating disturbed responses to kainic and leaving cell-bound 10-12 kDa C-terminal in ectodomain aspartate, neophobia and deficiency exploratory transmembrane and cytoplasmatic domains (Selkoe, 1994). behavior were demonstrated in these mice. In the of by a-secretase, production In the absence cleavage brain, the observed neuropathological changes of the 13A4 is possible by the action of other peptide included necrosis, apoptosis and astrogliosis in the cellular named and uncharacterized proteinases, hippocampus, cortex and other areas. The data demon- P- Cellular studies in vitro with the clinical APP y-secretases. strate that incomplete or incorrect a-secretase pro- demonstrated that the a-secretase cleavage is mutants of APP results in severe neurotoxicity and that cessing in favor of ,- and y-cleavage: considerably more reduced this effect is in a dominant manner. expressed is from the Swedish APP mutant peptide produced precursor protein/apoptosis/neuro- Keywords: amyloid OA4 et al., 1992) while the APP717 mutants (swAPP) (Citron mice degeneration/transgenic more of the longer 42-43 amino acid yield relatively PA4 et The Swedish mutation peptides (Iwatsubo al., 1994). also affects intracellular routing and processing of APP MDCK cells et al., 1994; De Strooper in polarized (Lo Introduction Haass et et al., 1995a; al., 1995). The function of amyloid precursor protein (APP) Whereas the known clinical mutations in APP affect only is the inhibition of conclusively demonstrated to date and no natural mutations have the y-secretase sites, P- for the molecu- with specificity at the a-secretase site. Located most closely serine proteinases arginine been described inhibitor domain. lar forms of APP that contain the Kunitz the clinical mutations Ala692-Gly and Glu693->Gln, are the P1 The accumulated molecular evidence complicates downstream of the a-secretase residue, Lys687. situated APP located for from the gene mutation is clinically responsible analysis considerably: single The Glu693->Gln in some cases com- at least seven isoforms can be genera- recurrent on chromosome 21, hemorrhages (HCHWA-D) The alternative mRNA (Sandbrink et al., with AD et Ala692-*Gly ted by splicing bined (Levy al., 1990). mixed clinical of AD and references the isoforms in a picture 1994 and therein). Moreover, mutation, resulting et was shown as transmembrane that can be recurrent hemorrhages (Hendriks al., 1992), are synthesized proteins secretion of APP et N- and to cause reduced (Haass al., 1994). processed, O-glycosyla- differently proteolytically of the a-secretase as well as all to To examine the importance processing sulfated, varying ted, phosphorylated, introduced a double mutation in a end of we have This heterogeneous product step, previously degrees. yields very two basic residues that is secreted into the the by replacing kDa ectodomain ca-secretase region a 110-130 © Press Oxford University D.Moechars et al. Fig. 1. Secretion of APP/WT and APP/RK in COS cells, polarized MDCK cells and cultures of rat neurons. Cells primary hippocampal were metabolically labeled and APP was from cell extracts immunoprecipitated labeled and from the conditioned medium (lanes CELL) (lanes labeled SUP) and analyzed by SDS-PAGE. In MDCK cells, precursor APP and the mature N- and APP can be O-glycosylated (APPm) (APPp) clearly separated, while the APP75 1 is indicated endogenous by asterisks. with two acidic residues. The APP/RK mutant resulting Table I. Overview of all founder mice with the APP/RK mutant was secreted from COS cells at a rate that was <50% cDNA relative to APP wild-type et (APP/WT) (De Strooper al., 1993). We here Founder No. report that, also from MDCK Copy Offspring Expression polarized (%) the secretion of the APP/RK cells, mutant is markedly a/RK/B/1 3 + <5 reduced without the affecting exclusively basolateral secre- a/RK/B/2 + <5 tion In use of a viral pattern. addition, making infection a/RK/B/3 -- system, we demonstrate that rat neurons a/RK/F/4 + primary secrete 3 a/RK/F/5 2 + <5 less ectodomain from than considerably APP/RK from a/RK/F/6 1 + APP/WT. The functional of the APP/RK repercussions a/RK/F/7 + 2 5 mutant were in mice analyzed transgenic by expressing ae/RK/B/I 5 + <5 the APP/RK cDNA the elements of using neuron-specific ae/RK/B/2 3 + <5 the mouse ae/RK/F/3 3 Thy-i gene promoter et The + (Vidal al., 1990). ae/RK/F/4 4 mice resulting transgenic displayed a disturbed behavior, ae/RK/F/5 2 which deteriorated progressively with and resulted in age ae/RK/F/6 15 + <5 premature as soon as 2 death, occurring months after birth ae/RK/F/7 20 + 5 for individual of some ae/RK/F/8 homozygotes transgenic lines. 10 + ae/RK/F/9 The 1 + results of behavioral molecular, biochemical, and ae/RK/F/10 2 + histochemical tests and are and experiments presented t/RK/F/l -- discussed. t/RK/B/2 1 + 50 t/RK/F/3 1 + 5 t/RK/F/4 + 1 45 Results t/RK/F/5 1 + 30 t/RK/F/6 2 + 10 Expression of APPIRK in COS cells, polarized MDCK cells and neurons primary COS cells were transiently transfected with constructs the containing cDNA coding for APP/RK or APP/WT cultures of rat hippocampal neurons were infected with under the control of the early SV40 promoter recombinant (De Strooper Semliki forest virus carrying the APP/WT or et al., 1993). Secretion of APP/RK was ~50% relative the APP/RK mutant cDNA. The secretion of APP/RK was to APP/WT, confirming our previous observations (De again <50% relative to secretion of APP/WT (Figure 1). et Strooper al., 1993). MDCK cell lines were created that have either APP/ Generation of APPIRK transgenic mice and WT or APP/RK cDNA driven by the RSV promoter stably analysis of expression integrated in their genome (De Strooper et al., 1995a). All transgenic lines derived using APP/RK cDNA in Three independent cell lines of each were grown polarized different constructs are summarized in Table I. Transgenic on filters, and the apical and basolateral compartments founders are denoted by a code identifying the promoter were analyzed for APP secretion. In standard conditions, (t, a or ae, respectively for the mouse Thy-i gene promoter, secretion of mutant APP/RK is considerably decreased the human APP gene promoter and the latter with a viral relative to APP/WT and remains exclusively in the baso- enhancer element), the APP cDNA type or mutant (in the lateral compartment (Figure 1), a situation not different present work only APP/RK), the mouse strain (F or B, from APP/WT (De Strooper et al., 1995a). Primary respectively for FVB or C57B1) and the line number. The 1266 Mice transgenic for a-secretase APP mutant mRNA in the wild-type and t/RK strains, with the larger mRNA size due to the 3'-untranslated region of the Thy-i gene. Expression by the human APP gene promoter was low relative to endogenous APP. The mouse Thy-i gene promoter resulted in brain expression levels in hetero- zygotes that reached 50% of endogenous APP mRNA levels (Table I, Figure 2). In homozygotes, APP/RK mRNA in brain approached those of endogenous levels APP. in the t/RK lines Restriction of expression to brain was demonstrated by RT-PCR and Northern blotting. Endogenous Thy-i gene was expressed in brain and thymus (Figure 2). The thymus-specific elements of the mouse Thy-i gene are present in intron 3 and deleted in the construct used (Vidal et al., 1990). Western blotting of brain extracts with different poly- clonal and monoclonal antibodies detected consistently more of the 110-130 kDa APP present in the brain of transgenic APP/RK mice relative to non-transgenic control mice (Figure 3A). Individual variations between mice were apparent but, overall, less important than the differences between the different transgenic strains or between mice with a different genetic status: heterozygotes expressed APP controls and levels that are intermediate between 3E). Semi- homozygotes of the same strain (Figure quantitative comparison indicated that brain APP levels in the heterozygotes were ~75% (line t/RK/F/6) and 200% The highest level (t/RK/F/4 strain) higher than in controls. of was noted in the brain of homozygous expression t/RK/F/4 mice 3C) amounting to ~3-fold the (Figure APP levels. These protein data largely endogenous paralleled the expression at the mRNA level. In all the APP/RK lines a larger APP protein band was detected that is hardly visible in control brain extracts (Figure 3). This uncleaved band, representing mature, proteolytically APP/RK is being characterized and its metabolism examined extensively, not only in the cellular paradigms in used above, but also primary cortical and hippocampal neuronal cultures derived from the transgenic mice. Demonstration of reactive gliosis in the transgenic mice us to the levels of (see below) prompted compare glial acidic Western fibrillary protein (GFAP) by blotting. Some increase in GFAP was in the evident, especially animals, but was not nearly as prominent as homozygous the increase in APP Combined immuno- (Figure 3). precipitation and Western blotting failed to detect sig- nificant amounts of the peptide in brain extracts from PA4 not normal and APP/RK transgenic mice (results shown). Fig. 2. Analysis of mRNA expression. (A) Northem blotting of total in brain was The expression of other genes the analyzed brain RNA from a non-transgenic and from five t/RK heterozygous mRNA the APP-related at the protein or level, including and from transgenic mice. (B) Northern blotting of RNA from brain APLP1 and the E proteins APLP2, apolipoprotein gene mouse. seven different tissues of a homozygous t/RKlB/2 transgenic 4 kb detected A2MRILRP and A2MRAP A mouse Thy-i gene exon probe (1.4 ApaI fragment) and its receptors VLDLR, kb I mRNA and the 3.9 kb mRNA. RT-PCR the 1.85 Thy- APP/RK (C) et The APP/RK strains did (Lorent al., 1995). transgenic brain mRNA PCR for APP/RK mRNA on using primers specific differ in this from not appreciably respect age-matched or APP mRNA (lanes primers L921-L57) endogenous (lanes 2, 1, control mice. Positive control was RNA from COS cells primers L922-L57). APP/WT or APP/RK transiently expressing (see Figure 1). of APPIRK Phenotypic characterization transgenic mice concern and homo- experiments mainly heterozygous death. Mice from the C57B1 and CBA/ of founders t/RK/F/ Premature FVB, zygous offspring t/RK/B/2, t/RKIF/4, As more strains live >2 6 and ae/RK/F/7 J normally years. progressively (Table I). this was of the in the APP/RK mice died RT-PCR established expression transgene prematurely, phenomenon a of 64 in four monitored in in four out of 10 ae/RK population heterozygous transgenic brain transgenic strains, relative to a of 30 in five out of six t/RK lines mice of the first out of six a/RK lines and generation, group littermates. In the first revealed 3.5-3.9 kb Northern genotyped, non-transgenic year, (Table I, Figure 2). blotting 1267 D.Moechars et al. only 7% of non-transgenic mice while 66 mild seizures of died, 44, and transfer varying severity and, upon 69% of transgenic mice died in lines into clean t/RK/F/6, t/RK/F/4 cages, posture and Sex freezing neophobia. and t/RKJB/2, respectively (Figure 4). The first death difference was not Seizures in consisted of facial apparent. lines tIRKIB/2, t/RK/F/4 and t/RK/F/6 occurred at mild fore-limb clonus and 30, 78 movements, mild whole body and 150 days of clonus age, respectively. This correlated from to 10 s. with lasting S Transgenic mice older than the higher 6 APP/RK expression level in lines months in severe t/RKlB/2 and showed, addition, tonic-clonic seizures, t/RK/F/4 relative to t/RK/F/6 30-60 (Table I). lasting s. These mice and stretched gasped all four All animals that died were limbs with spontaneously weighed and their tail followed upright, by severe whole necropsied, but no malformations of the vital clonic organs nor seizures. mice were body Afterwards, for lethargic major internal bleeding were observed. Premature variable of death periods time. Seizures have been observed in was not preceded by overt of deteriorated health >60% of all signs or mice in these two lines. transgenic Seizures by marked loss of weight. were absent or rare in mice of line t/RK/B/2 very (Table which were I), backcrossed into the C57Black background In Spontaneous behavior lines t/RK/F4 and t/RK/F/6, were at (mice analyzed least 87.5% of this background). heterozygotes of <3 months and of <2 homozygotes This strain is known to be more resistant to seizures and months behaved normally. showed Subsequently, they to drugs inducing seizures (Ganesan et 1995 and al., increased agitation, bouts of wild running and, among references therein). Spontaneous seizures were never more males, frequent and fierce fighting. Many displayed observed in normal mice of the FVB strain in our which are under colony, kept identical conditions to the transgenic mice. Behavioral In a new testing. environment, normal mice move immediately towards the wall of the cage and the walls proceed along (thigmotaxis) (Walsh and Ambulation as an Cummins, 1976). indicator of motor was activity scored of by parameters latency, i.e. the time elapsed before the beginning of thigmotaxis and by the number of corners crossed during the 30 s interval. In lines t/RK/F/4 and tIRK/F/6, respectively, 54 and 47% of the mice exhibited a latency of >5 as to 5% s, opposed of control FVB mice. Younger transgenic mice behaved like age-matched littermates: at 4-6 weeks, <10% of mice of both lines transgenic displayed a latency of >5 s. Less than 2% of FVB controls showed posture freezing in a new environment, as opposed to 30% of t/RK/F/4 and 23% of t/RK/F/6 mice (3-5 months old). Transgenic mice of all ages also showed a markedly increased respiratory rate, while some mice went into seizures, indicating the severe stress. In a group of 35 FVB control mice, the frequency of corner crossing was Gaussian distributed, centered around a mean of seven crosses (Figure 5). This distribution was Fig. 3. Western blotting of APP and GFAP in brain of control and transgenic mice. (A) FVB control mice (lanes 1-3) and homozygous t/RK/F/4 transgenic mice (lanes 4-6) immunoblotted with mAb 22C1 1 (Hilbich et al., 1993) on a 6% homogeneous SDS-PAGE gel. (B) The same extracts separated on a 4-20% linear SDS-PAGE gradient gel, immunoblotted with a polyclonal antibody against GFAP. (C) Different amounts of brain extract of FVB control (extract 1 from A) and homozygous t/RK/F/4 (sample 4 in A) on a 6% gel, stained with mAb 22C11. Lane 1: 6 lane 2: 2 lane 3: 0.6 ,l, lane 4: 0.2 p1, gA, gl, lane 5: 0.2 lane 6: 0.6 pl, lane 7: 2 Note the equivalent staining gl, g1. intensity of lanes 3 and 5 (volume extract ratio of 3) and the stronger staining of the larger extra band of APP in lanes 5-7. (D) Brain extracts of different control and heterozygous transgenic mice, separated on a 4-20% gradient gel and stained with mAb 22C Lane 1: FVB control, lane 2: Fl(C57BlXCBA/J) control, lane 3: heterozygote t/RK/B/2, lane 4: heterozygote t/RK/F/4, lane 5: heterozygote t/RK/B/6, lane 6: heterozygote of a mouse, t/APP770/F/1 transgenic for mouse APP770 cDNA driven by the mouse Thy-l gene promoter. (E) Brain extracts of different heterozygous and homozygous transgenic mice separated on a 4-20% gradient gel and stained with mAb 22C11. Lanes 1 and 2: heterozygous and homozygous t/RK/F/4, lanes 3 and 4: heterozygous and homozygous t/RK/F/6, lanes 5 and 6: heterozygous and homozygous t/RK/B/2, lane 7: FVB control. 1268 transgenic for a-secretase APP mutant Mice 80. 25 - a) 20 - M 70 15 - I'2 8) 5- .I.I.I.I. I. AL X <~~~~~~~~~~5. -zz---x 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 =30 control n .> 30 13 -4-I-t/RK/B/2 n =10 -X-tURKIFI3 CO 20 -?a-I n 18 VRK/F/4 10 -X-tVRK/F/6 n 18 -C-O-tRKIF/4 Ho n =22 0 I1 1 1 1 1 1 1 I 0 0 0 0 0 0 0 0 0 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 (D) 0 03) (a LO co 0 C _N c' cn c' c Age (days) Fig. 4. Premature death of control FVB, heterozygous and 25- homozygous APP/RK mice. The death rate in four t/RK lines that were heterozygous for the APP/RK transgene compared with non- 15- transgenic littermates over a 1 year period. Homozygous offspring of the t/RK/F/4 line were followed for only 105 days. 4 5 6 7 8 9 10 11 12 13 14 0 1 2 3 15, 0 1 2 3 4 S 6 7 a 9 10 11 12 13 14 15 similar for young or old FVB mice (2-3 or 9-12 months Corners crossed old). Heterozygous and homozygous mice of lines t/RK/ t/RK/F/6 (6-20 weeks old) displayed significantly F/4 and ambulation. Strikingly, the frequency of corner reduced crossings was skewed towards low values, separating two one with low explorative behavior and subpopulations, with only a slight reduction. Afterwards, the tested another 150 animals did not behave differently from untested transgenic 5 120-- mice. It was also evident that premature death was not ~'0 confined to either subpopulation of the tested transgenic mice: two out of five t/RKJF/4 mice and two out of M 30 -- 11 t/RK/F/6 mice (from those tested in Figure 5) that died ni during the 6 months subsequent to the open field test, Test I Test belonged to the low explorative groups. This proved that the reduced exploratory behavior was not a direct sign of death of the transgenic animals. imminent, premature aggressive behavior in the t/RK/F/4 The augmented 30 () 25 offspring was measured under controlled and t/RK/F/6 o 20 Single male FVB control and transgenic mice, conditions. for the transgene, were housed isolated for homozygous <10 4 before being confronted in their cage 5 weeks ('resident') with a male FVB 'intruder' that had been reared in a was scored by the latency of the group. Aggression I Test 2 Test first attack and by the number of attacks during 3 min 5. Behavioral testing. (A) In open field testing, the number of (Saudou et al., 1994). APP/RK resident males Fig. observation crossed during a 30 s interval was measured. FVB control corners attacked sooner and more often than non-transgenic FVB mice panel, n = 35), t/RK/F/4 mice (middle panel, n = 30) (upper resident males. Transgenic males from strain t/RK/F/4 n = (B) In the aggression test, mice (lower panel, 50). and t/RK/F/6 attacked sooner and more often than t/RK/F/6 residents were FVB bars, n = 4), homozygotes of lines resident mice (black the attacks were judged to be fiercer and n = and t/RK/F/6 (striped bars, n = 5). 5), and t/RK/F/4 (open bars, 6) (Figure attack panel) and the number of attacks by the Latency of (upper intense. When the same animals were retested, more in a 3 min observation period. Data are resident (lower panel) was increased (Figure 5). aggressiveness with standard errors. Statistical analysis mean values calculated between and control mice to be showed the difference transgenic to glutamate analogs Reactivity (P<0.02). significant The system is the major excitatory system glutamatergic in the generation of seizures. Its activity was involved for KA induction of seizures and curves Dose-response tested in lines t/RKIF/4 and t/RK/F/6 with two agonists and mice in C57B1 FI(C57BlXCBA/J) mortality FVB, of different glutamate receptors, i.e. kainic acid (KA) and strain differences being more obvious (FVB demonstrated (NMDA). N-methyl-D-aspartate 1269 D.Moechars et al. hilus of the dentate gyrus and dispersed throughout the 80 - granular layer of the dentate gyrus. In the most affected 60 - animals, neurons were in the degenerating present thalamus 40 - (Figure 7) and, to a lesser extent, in the posterior cingulate 20- cortex, in the caudate putamen, the and hypothalamus the CAI and CA2 region of the 0 hippocampal pyramidal layer. c 32 Typical also were the nuclei of glial cells in the 24 40 corpus callosum and in the fimbria which are normally ellipsoid, Dose of KA (mg/kg) 0 but appeared rounded and more densely stained in APP/ RK mice (Figure 7). The most severe signs of neuro- degeneration were encountered in four t/RK/B/2 and in two t/RK/F/4 mice. The age of the animals was not the determining factor, since these mice were aged between 2.5 and 9 months. The abnormal neurons in the dentate gyrus were reminis- cent of pyknotic or apoptotic cells: condensation of the chromatin results in compact and heavily stained nuclei 40 50 60 80 100 with blebbing of the nuclear membrane. Apoptotic cells Dose of NMDA (mg/kg) have been demonstrated by immunochemical methods in AD brain (Su et al., 1994) and in transgenic mice (LaFerla Fig. 6. Effects of glutamate analogs kainic acid and NMDA. Upper of KA to FVB mice et al., 1994). In situ detection of apoptotic cells in brain panel: mortality resulting from i.p. administration = mice (black bars, n 4, 9 and 6), homozygous t/RK/F/6 bars (open slices by terminal transferase incorporation of digoxigenin- = mice n = n 2, 8 and 4) and homozygous t/RK/F/4 (striped bars, 2, demonstrated dUTP and peroxidase-based detection, con- from of 9 and 4). Lower panel: mortality resulting i.p. injection siderable numbers of apoptotic cells in t/RK/B/2, t/RK/ in n = 4 and NMDA FVB mice (black bars, 4, 6, 7, 2), homozygous mice In mice of line t/RK/ F/4 and ae/RK/F/7 (Figure 7). n = 4 and and t/RK/F/6 mice (open bars, 2, 4, 4, 3) homozygous mice = 6 and of line B/2 and some of line the of the t/RK/F/4 (n 4, 6, 5, 2). Homozygous offspring t/RK/F/4, granular layer were and and died when t/RK/F/4 resistant, displayed symptoms only dentate contained while in gyrus many apoptotic neurons, receiving higher doses (see text). ae/RK/F/7 mice apoptotic neurons were in the present amygdala (Figure 7). sensitive than but KA Reactive gliosis or astrogliosis was demonstrated by C57Bl mice), consistently elicited, dose clonic seizures which were rated in five immunostaining for GFAP, and proved positive even in dependently, stages: immobilization, head forelimb younger transgenic mice and in mice from line t/RK/B/2 nodding, clonus, rearing and rearing with falling (Racine, 1972). All APP/ that did not exhibit seizures. An intense astrocytic reaction KA was in RK mice were more sensitive to than were control observed the cortex, the amygdala and the hippo- KA seizures of mice (Figure 6). at 40 mg/kg triggered campus; astrocytes with enlarged cell bodies and thicker 5 while stained for Brains stage in all t/RK/F/4 and t/RK/F/6 mice, only processes intensely GFAP (Figure 7). of FVB control mice showed 4 seizures and of mice from lines and 66% stage transgenic t/RKIF/4 t/RK/F/6 were 30% 5. the all t/RK/F/4 mice stained for the with five stage Following seizures, immunochemically peptide PA4 died, while 35% of t/RK/F/6 mice and 50% of control different antibodies. some was evident in Although staining FVB mice survived (Figure 6). certain neurons, the reaction was weak not (results shown). The for NMDA in FVB control mice was LD50 ~50 mg/kg body weight (Figure 6). The APP/RK trans- Discussion mice genic demonstrated a markedly diminished sensitivity for NMDA. mice even resisted a Homozygous t/RK/F/4 The functional importance of correct a-secretase pro- dose of 100 which was of APP was mg/kg, invariably lethal for control cessing examined by creating a mutant APP FVB as well as for all t/RK/F/6 mice that was transgenic (Figure partially defective in secretion from three different 6). In t/RK/F/4 mice this dose of NMDA failed to cell COS high types: cells, MDCK cells and primary neurons. elicit the typical NMDA symptoms evident in control Expression of this a-secretase mutant in mouse brain FVB mice (Leander et al., 1988). These symptoms were resulted in a progressive disorganization of the central evident at 150 a dose mg/kg, that was never tested in nervous system, resulting in behavioral disturbances, control or t/RK/F/6 mice and which killed all t/RK/F/4 seizures, differential reaction to glutamate analogs, prema- transgenic mice in 20 min. ture death and morphological neuronal changes in the brain. The histological analysis demonstrated conclusively Histological analysis of the brain of transgenic that the brain of transgenic t/RK mice contained many APPIRK mice abnormal neurons in the dentate and in the thalamus gyrus In the brain of 24 out of 25 transgenic mice of lines and, in older mice, also in the cortex and the pyramidal t/RK/B/2 (four mice), t/RK/F/4 (15 mice), t/RK/F/6 (three layer of the hippocampus. The symptoms of deranged mice) and ae/RK/F/7 (three mice) that died during observa- behavior, the occurrence of seizures and the premature tions, typical abnormal neurons were evident: many death are a consequence of the documented neuronal neurons were vacuolized or contained clear cytoplasm, damage. This conclusion is derived from the accumulated surrounding compacted and densely stained nuclei (Figure data in presented the Results section and discussed below. 7). These aberrant neurons were most conspicuously Expression of the transgene was restricted exclusively located on the boundary of the granular layer and the to the brain, as anticipated from the characteristics of the 1270 Mice transgenic for a-secretase APP mutant A B * 0 S *1'6* 9% ~~~~~~~~~~~4 0409 ~~~~~0 4 40* S. ... ,s:t ;, 4% 94 h.~ t 'IA *ee 01.~~~~ 49~~~~~~~ E F J H a0I 4F .S . i i':l Oeb * e St ",W. _.,,. ' X # .! k * v , 2 * # s . * it 0 41k... -1*it * .> \ .. ... t.00 ;. 4 ..s s; S 0 A.tt 0 * oe K L :thi ~ ~ : ~ ~ ?. p S :V .A- :lk S, ,A rk A; * 7. and GFAP of mouse brain. Fig. Histology, apoptosis immunochemistry of sections of the brain of FVB (A-J) Hematoxylin-eosin staining sagittal and APP/RK mice. show the of FVB Representative photomicrographs granular the dentate from control mouse t/RK/B/2 mouse layer gyrus (A), (B), t/RKIF/4 mouse and t/RK/F/6 mouse of (C) (D); the thalamus from control FVB mouse and t/RK/F/6 mouse of the callosum (E) (F); corpus from control FVB mouse and t/RK/F/6 mouse of the CAl from control FVB mouse and t/RK/B/2 mouse (G) (H); hippocampal pyramidal layer (I) (J) (X60). (K-N) In situ apoptosis in the brain of APP/RK mice detected on sections in the dentate from control FVB transgenic sagittal gyrus (K), t/RK/B/2 and t/RK/F/4 mice and in the of an mouse GFAP in the cortex of control (L) (M) amygdala ae/RK/F/7 (N) (x 100). (O-S) immunostaining a of the FVB (0) and t/RK/F/4 mice in the dentate of FVB and mice is (P) (x 10), gyrus (Q) t/RK/F/6 (R) (X20). (S) higher magnification (X40) bottom right hand corner of (R). the mouse used in the mRNA and level. the elements of at protein Thy-i gene promoter Although homozygous the most severe and the constructs et In the four of strain t/RK/F/4 injected (Vidal offspring display al., 1990). mild as in mouse strains and even independent transgenic spectacular phenotype, analyzed overexpression, resulted in the of the was related of line that described, severity homozygotes t/RKIF/6, symptoms phenotype same. In not to the of the in the brain both are the directly expression essentially addition, although presented transgene 1271 D.Moechars et al. processed a-secretase. here, a similar phenotype is becoming evident in by Obviously, this is unique since transgenic the APP/RK mice that express the same APP/RK mutant but mutant is artificial and it is impossible to cDNA, how the double under the control of the human APP gene promoter predict mutation at the C-terminal end (the a/RK and The in these strains of the secreted APP/RK ae/RK lines). symptoms are ectodomain would disturb an milder and at a later which is consistent function. develop age, with (unknown) Alternatively, the slower processing with a-secretase the lower expression level attained this by should favor [-cleavage of promoter APP/RK, in (unpublished results). Further, we have generated trans- resulting secretion of [3-cleaved APP/RK, a situation to genic mice in which the mouse be in Thy-i gene promoter envisaged patients carrying the Swedish APP muta- elements are used to drive of the tion et expression mouse (Citron al., 1992). APP770 cDNA or of the C-terminal domain These considerations should take (Figure 3) of into account the APP fused to its signal These intracellular fate of peptide (not shown). unprocessed APP/RK and of cell- mice show none of the bound the transgenic phenotypic character- fragments containing [A4 peptide resulting istics of the APP/RK mice. We conclude that the observed from [-cleaved APP. An interesting collateral is provided phenotype is specific and caused the over- by the recent report of transgenic mice directly by that overexpress expression of the proven a-secretase APP/RK mutant in the [BA4 peptide intracellularly et (LaFerla al., 1995). The the brain. many phenotypic traits that are similar to those observed Some of the of the APP/RK in the APP/RK mice described here invite phenotypic parameters speculation mice are reminiscent of i.e. dis- about a common cause. Possibly, the cell-bound AD, neurodegeneration, APP/RK turbances in behavior and premature death. C-terminal fragments can lead to intracellular Preliminary [A4 peptide, tests indicated that the short-term of the creating a situation that is not essentially different from learning ability the is one mice not affected, but more extensive and careful in which the [A4 peptide is produced intracellularly as testing of and is in learning memory capabilities progress. The instructed by the transgene (LaFerla et al., 1995). This occurrence of seizures in our generalized transgenic mice reasoning implies that intracellularly produced [3A4 pep- was but literature data that these are tide could unexpected, prove cause neurodegeneration and, eventually, could frequently observed in AD even as a be more neurotoxic than patients, prominent extracellularly produced [A4 feature in familial chromosome 14-linked while less peptide. This AD, would explain why transgenic mice with a prevalent in with the APP717 mutations massive 10-fold patients (Mullan overexpression in brain of a clinical APP et al., 1993; Lampe et It was debated whether mutant protein on the one al., 1994). displayed, hand, massive seizures constitute an characteristic of AD or deposits of [-amyloid in authentic integral extracellular amyloid are a secondary phenomenon. The loss of neurons and plaques but, on the other, produced a only very mild important alterations of the architecture of phenotype (Games et al., 1995). more hippocampal Clearly, extensive and neocortical areas provides a suitable substrate for the comparative and longitudinal studies are on required the generation of seizures, while seizures could enhance the different transgenic mouse strains. of the disease. progression An observation that could be related to directly the Glutamate has been in mechanisms involved as excitotoxicity implicated the discussed is the production of neurodegeneration and in neuronal cell in or [3A4 process death amyloid peptide in mice. Little or no [BA4 peptide- AD. In the tests with the glutamate the observed are in analogs, containing peptides produced COS cells, MDCK refractiveness to NMDA of the APP/RK mice cells or neurons after transfection transgenic with mouse APP as could be 'NMDA indicative of opposed to human APP et receptor hypofunction' (De Strooper al., 1995b and recently as a proposed novel type of excitotoxicity unpublished results). On the other hand, intracellular (Rothman and In Olney, 1995). addition, since NMDA amyloid-containing fragments are formed readily from receptor activity is to be mouse thought linked intimately to the APP695 after 'humanization' by changing only central mechanisms of and three amino memory building learning acid residues in the [BA4 peptide region of (Malenka, 1994; Collingridge and Bliss, we believe murine 1995), APP695 (De Strooper et al., 1995b). The fact that that the APP/RK mice will allow further studies in we used mouse APP this to generate the mutant APP/RK direction. The difference in reactivity in this mice respect, transgenic contributed to or highlighted precisely caused by the APP/RK transgene in the different genetic that aspect that otherwise might have been blurred, i.e. backgrounds (FVB or C57Black, respectively in the t/RK/ that the normal and pathological physiology of APP in F/4 or t/RK/F/6 and t/RKIB/2 transgenic lines), is most brain includes more than production of the [A4 peptide. interesting: the major and so far only difference is the In this respect, it is noteworthy that APP-deficient mice almost complete absence of seizures in line t/RK/B2 while, demonstrated several phenotypic characteristics in com- by the other parameters and criteria of molecular and mon with the APP/RK transgenic mice (Zheng et al., histochemical analysis, this line is not notably different 1995 and unpublished results). Eventually, not only the from the FVB transgenic lines. The C57Black mouse production of peptide but also the absence or suppres- PA4 strain is known to be more resistant to seizures and to drugs sion of an as yet unknown function of APP will explain inducing seizures (Ganesan et al., 1995 and references this paradox. therein). This clear genetic component in the phenotype is another strong argument that the seizures are the consequence and not the cause of the neurodegeneration. Materials and methods To explain the effect of the APP/RK mutation molecu- Construction of the APP695/RK minigenes larly, two possibilities need to be considered. Given the The mouse APP/RK cDNA (De Strooper et al., 1993) was cloned in an location of the double mutation, the secreted APP/RK adapted site in the vector (gift SfiI pTSCaxI from H.Van der Putten, ectodomain will contain a mutated C-terminal end when Basel). This pUC18-based vector contained an 8.1 kb EcoRI fragment 1272 Mice transgenic for a-secretase APP mutant comprising the mouse Thv-1.2 gene (Vidal et al., 1990). A 1.5 kb BanI- (Ciba-Geigy, Basel), C.Dotti (EMBL, Heidelberg), B.Greenberg XhoI fragment (located in exon 2 and exon 4, respectively) was replaced (Cephalon, West Chester, PA), M.Crauwels (Leuven) and S.Dunnett by the APP/RK cDNA. Alternatively, the mouse cDNA was (Cambridge). APP/RK Expert technical assistance was provided by M.Gilis, cloned downstream of the human APP gene promoter with a ,B-globin C.Kuiperi, K.Craessaerts, S.Torrekens and L.Stas. This investigation intron interspersed. A duplex polyoma virus enhancer element eventually was supported by the 'Fonds voor Geneeskundig Wetenschappelijk preceded the human APP promoter. Onderzoek' (FGWO-NFWO), the Interuniversity network for Funda- mental Research (IUAP, 1991-1996) and the Action Program for Microinjection and analysis of founders Biotechnology of the Flemish government (VLAB, ETC-008). B.D.S. is Linearized constructs were purified and microinjected into 1.5-day-old a postdoctoral fellow of the National Fonds voor Wetenschappelijk pre-nuclear embryos taken from superovulated FVB or Fl (C57BIX Onderzoek (NFWO) and thanks the Sandoz Foundation and the Divry CBA/J) females. The injected oocytes were allowed overnight to reach Research Fund for support. D.M. is the recipient of an IVT/IWONL the two-cell stage and transferred into pseudopregnant fosters, Fl scholarship. (C75BIXCBA/J). Tail biopsies were taken 3 weeks after birth at weaning for isolation of genomic DNA and Southern blotting, as described References (Umans et al., 1994). Chartier-Harlin,M.C. et al. (1991) Ea-rly-onset Alzheimer's disease Analysis of RNA and protein caused by mutations at codon 717 of the precursor protein 0-amyloid Total RNA from mouse tissues was isolated and analyzed by Northern gene. Nature, 353, 844-846. blotting as described (Lorent et al., 1995). RT-PCR was carried out with Citron,M., Oltersdorf,T., Haass,C., McConlogue,L., Hung,A.Y., sense primers containing either the wild-type APP or the mutated APP/ Seubert,P., Vigo-Pelfrey,C., Lieberburg,I. and Selkoe,D.J. (1992) RK region (De Strooper et al., 1993): primer L921, 5'ttgaagtcCGCcat- Mutation of the ,B-amyloid precursor protein in familial Alzheimer's caaAaA3' or primer L922, 5'ttgaagtcGATcatcaaGaG3' (the five nucleo- Disease increases :-protein production. Nature, 360, 672-674. tide differences are capitalized). The antisense primer was L57: 5' Collingridge,G. and Bliss,T. (1995) Memories of NMDA receptors and ccgatgggtagtgaagcaatggtt 3'. RT-PCR was done with touch-down LTP. Trends Neurosci., 18, 54-56. temperature programming: four series of three cycles each (1 min De Strooper,B., Umans,L., Van Leuven,F. and Van Den Berghe,H. (1993) denaturation at 92°C, 2 min annealing at 68, 67, 66 and 1 65°C and min Study of the synthesis and secretion of normal and artificial mutants extension at 72°C) followed by 35 cycles 92°C, 2 (I min min 64°C, of murine amyloid precursor protein (APP): cleavage of APP occurs 1 min 72°C). For Western blotting, brain tissue was homogenized in in a late compartment of the default secretion pathway. J. Cell Biol., vol of buffer (50 mM Tris, pH 8.0, 150 mM NaCl, 1% deoxycholate, 121, 295-304. 1% Triton X-100, 0.1% SDS) with proteinase inhibitors (25 mM EDTA, De Strooper,B., Craessaerts,K., Dewachter,I., Moechars,D., ,ug/ml pepstatin, 1000 U/ml aprotinin) in a motor-driven mechanical Greenberg,B., Van Leuven,F. and Van Den Berghe,H. (1995a) homogenizer. After centrifugation (12 000 g, 20 the clear min), super- Missorting of amyloid precursor protein in MDCK cells. J. Biol. natant was made 1% in SDS and 2-mercaptoethanol, boiled for 5 min Chem., 270, 4058-4065. and analyzed on Tris-glycine-buffered polyacrylamide gels (6% or De Strooper,B., Simons,M., Multhaup,G., Van Leuven,F., Beyreuther,K. 4-20%). and Dotti,C. (1995b) Production of intracellular amyloid containing fragments in hippocampal neurons expressing mutants of human Brain histology and immunostaining amyloid precursor protein causing Alzheimer's disease and protection Brain tissue was fixed ovemight in 4% paraformaldehyde, dehydrated against amyloidogenesis by subtle amino acid substitutions in the and paraffin embedded for sectioning (5 ,um). For in situ apoptosis rodent sequence. EMBO J., 14, 4932-4938. detection, sections were mounted on silanized slides and treated for Games,D. et al. (1995) Alzheimer-type neuropathology in transgenic terminal transferase incorporation of digoxigenin-dUTP and immuno- mice overexpressing V717F ,B-amyloid precursor protein. Nature, 373, peroxidase detection and staining (Apoptag kit, Oncor). For GFAP 523-527. immunostaining, 5 paraffin sections were dewaxed, rehydrated and Ganesan,L., Kofuji,P., Wang,J., Fernando,J., Liu,Z., Mathura,J. and gm treated with hydrogen peroxide to eliminate endogenous peroxidase Burt,D. (1995) GABA receptor ,B2 and ,B3 subunits: comparison in PI, activity. After blocking with normal goat serum (5% in 50 mM Tris, DBA/2J and mice. C57B1/6J Biochim. Biophys. Acta, 1261, 134-142. pH 7.4, 150 mM NaCl, 0.1% Triton X-100), sections were incubated Glenner,G.G. and Wong,C.W. Alzheimer's disease and Down's (1984) overnight at 4°C with rabbit anti-cow GFAP antibody (Dakopatts, syndrome: sharing of a unique cerebrovascular fibril amyloid protein. Denmark) diluted 1:1000, followed by peroxidase-conjugated goat anti- Biochem. Biophys. Res. Commun., 122, 1131-1135. rabbit antibody (dilution 1: 100). The reaction was developed with Goate,A. et al. (1991) Segregation of a missense mutation in the amyloid diaminobenzidine and hydrogen peroxide. precursor protein gene with familial Alzheimer's disease. Nature, 349, 704-706. Open field test Goldgaber,D., Lerman,M., McBride,O., Saffiot,U. and Gadjusek,D. All mice used were kept under a light cycle of 12 with h, food and (1987) Characterization and chromosomal localization of a cDNA water ad libitum. The corner index test is a variant open field test to of encoding brain an'yloid Alzheimer's disease. Science, 235, 877-880. measure activity and ambulation. The mice were placed in the middle Haass,C., Hung,A.Y., Selkoe,D.J. and Mutations Teplow,D.B. (1994) of a clean, unused opaque cage (20 cm square) and allowed to move associated with a locus for familial Alzheimer's disease result in under dim, indirect light for 30 s. Between tests, cages were cleaned. alternative processing of J. Biol. amyloid protein precursor. Chem., Controls were age matched and subjected at random to the tests. Results 269, 1-8. were scored independently by two observers, one of which was unaware Haass,C., Koo,E., Capell,A., Teplow,D. and Polarized Selkoe,D.J. (1995) of the genetic status of the mice. The mice were tested twice with a 3 sorting of 3-amyloid precursor protein and its in proteolytic products day interval. MDCK cells is regulated by two independent signals. J. Cell Biol., 128, 537-547. Aggression test Hendriks,L. et al. (1992) Presenile dementia and cerebral haemorrhage FVB or transgenic male 'residents' (tested at 10-12 weeks were old) linked to a of the mutation at codon 692 f-amyloid precursor protein housed alone for 4 weeks in normal cages in the same room under the gene. Nature 218-222. Genet., 1, same conditions as other mice. Bedding was changed once a week Hilbich,C., Monning,U., and Grund,C., Masters,C. Beyreuther,K. (1993) without handling the first the animal, with the last change 6 days before Amyloid-like of the of properties peptides flanking epitope amyloid in of test. 'Intruders' were male FVB mice (8 weeks old) kept groups precursor protein specific monoclonal 22C 11. J. Biol. antibody Chem., eight in larger After the intruder was in the resident cages. placed cage, 268, 26571-26577. attack number of attacks were recorded 3 min. An latency and during and Iwatsubo,T., Odaka,A., Suzuki,N., Mizusawa,H., Ihara,I. Nukina,N. attack when the resident bit the intruder. One week later was scored only of A beta and A beta 40 in senile (1994) Visualization 42(43) plaques the same resident mice were tested a second time with different intruders. that an with A beta monoclonals: evidence end-specific initially 45-53 deposited species is A beta 42(43) Neuron, 13, Kang,J., Lemaire,H.G., Salbaum,J.M., Masters,C.L., Unterbeck,A., Acknowledgements and Grzeschik,K.H., Muller-Hill,B. Multhaup,G., Beyreuther,K. Alzheimer's disease A4 The contribution or materials and the criticism of The of of expertise positive (1987) amyloid protein precursor 733-736. the following scientists is H.Van der Putten resembles a cell-surface 325, gratefully acknowledged: Nature, receptor. 1273 D.Moechars et al. and LaFerla,F., Tinkle,B., Bieberich,C., Haudenschild,C. Jay,G. (1995) The Alzheimer's peptide induces neurodegeneration and apoptotic AP cell death in transgenic mice. Nature Genet., 9, 21-29. of Lampe,T. et al. (1994) Phenotype chromosome 14-linked familial in Ann. Alzheimer's disease a large kindred. Neurol., 36, 368-320. Leander,J., Lawson,R., Ornstein,P. and Zimmerman,D. (1988) N-Methyl- in D-aspartic acid-induced lethality mice: selective antagonism by Brain phenylcyclidine-like drugs. Res., 448, 115-120. Levy,E., Carman,M.D., Femandez-Madrid,I.J., Power,M.D., Van Lieberburg,I., Duynen,S.G., Bots,G.T.A.M., Luyendijk,W. and Frangione,B. (1990) Mutation of the Alzheimer's disease amyloid in gene hereditary cerebral haemorrhage, Dutch type. Science, 248, 1124-1126. Lo,A., Haass,C., Wagner,S., Teplow,D. and Sisodia,S. (1994) Metabolism of the 'Swedish' variant in amyloid precursor protein Madin-Darby canine cells. 30966-30973. kidney J. Biol. Chem., 269, De Van Lorent,K., Overbergh,L., Moechars,D., Strooper,B., Leuven,F. and Van den in mouse and in Berghe,H. (1995) Expression embryos adult mouse brain of three members of the amyloid precursor protein of the related family, alpha-2-macroglobulin receptor/LDL receptor and of its protein ligands apolipoprotein E, lipoprotein lipase, alpha- and the 40 kDa associated 2-macroglobulin receptor protein. 1009-1025. Neuroscience, 65, Malenka,R. (1994) Synaptic plasticity in the hippocampus: LTP and LTD. 535-538. Cell, 78, and Masters,C., Simms,G., Weinman,N., Multhaup,G., McDonald,B. Beyreuther,K. (1985) Amyloid plaque core protein in Alzheimer disease and Down Proc. Natl Acad. Sci. 4245- syndrome. USA, 82, Mullan,M. et al. (1993) Clinical comparison of Alzheimer's disease in Ile in pedigrees with the codon 717 Val mutation the amyloid precursor protein gene. Neurobiol. Aging, 14, 407-419. Murrell,J., Farlow,M., Ghetti,B. and Benson,M.D. (1991) A mutation in the associated with amyloid precursor protein hereditary Alzheimer's Disease. 97-99. Science, 254, Modification of seizure Racine,R. (1972) activity by electrical stimulation: II. Motor seizure. Clin. Electroencephalogr Neurophysiol., 281-294. 32, and and the NMDA Rothman,S. Olney,J. (1995) Excitotoxicity receptor still lethal after Trends 57-58. eight years. Neurosci., 18, Rumble,B., Retallack,R., Hilbich,C., Simms,G., Multhaup,G., Martins,R., Hockey,A., Montgomey,P., Beyreuther,K. and Masters,C.L. (1989) A4 Amyloid protein and its precursors in Down's syndrome and Alzheimer's disease. New Engl. J. Med., 320, 1446-1452. and Sandbrink,R., Masters,C. Beyreuther,K. (1994) protein PA4-amyloid mRNA isoforms without exon 15 are precursor ubiquitously expressed in rat tissues but not in neurons. J. including brain, Biol. Chem., 269, 1510-1517. Ait Saudou,F., Amara,D., Dierich,A., LeMeur,M., Ramboz,S., Segu,L., and Enhanced behavior in mice Buhot,M. Hen,R. (1994) agressive lacking receptor. Science, 265, 1875-1878. 5-HTIB Cell of the and Selkoe,D. (1994) biology amyloid ,B-protein precursor the mechanism of Alzheimer's disease. Annu. Rev. Cell Biol., 10, 373-403. and Su,J., Anderson,A., Cummings,B. Cotman,C. (1994) Immunohisto- chemical evidence for in apoptosis Alzheimer's disease. NeuroReport, 2529-2533. 5, Tanzi,R.E. et al. (1987) Amyloid n-protein gene: cDNA, mRNA distribution and genetic linkage near the Alzheimer locus. Science, 235, 880-884. Umans,L., Serneels,L., Hilliker,C., Stas,L., Overbergh,L., De Strooper,B., Van and Van Leuven,F. den Berghe,H. (1994) Molecular cloning of the mouse for gene coding alpha-2-macroglobulin and targeting of the gene in embryonic stem cells. Genomics, 22, 519-529. Vidal,M., Morris,R., Grosveld,F. and Spanopoulou,E. (1990) Tissue- control specific elements of the Thy-] gene. EMBO J., 9, 833-840. Walsh,R. and Cummins,R. (1976) The open-field test: a critical review. 482-502. Psychol. Bull., 83, Zheng,H. et al. (1995) f-Amyloid precursor protein-deficient mice show reactive gliosis and decreased locomotor activity. Cell, 81, 525-531. Received on September 20, 1995; revised on November 7, 1995 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png The EMBO Journal Springer Journals

Expression in brain of amyloid precursor protein mutated in the alpha‐secretase site causes disturbed behavior, neuronal degeneration and premature death in transgenic mice.

Download PDF
10 pages

Loading next page...
 
/lp/springer-journals/expression-in-brain-of-amyloid-precursor-protein-mutated-in-the-alpha-ak0x2w115Y

References (40)

Publisher
Springer Journals
Copyright
Copyright © European Molecular Biology Organization 1996
ISSN
0261-4189
eISSN
1460-2075
DOI
10.1002/j.1460-2075.1996.tb00468.x
Publisher site
See Article on Publisher Site

Abstract

The EMBO Journal vol.15 no.6 pp.1265-1274, 1996 Expression in brain of amyloid precursor protein mutated in the a-secretase site causes disturbed behavior, neuronal degeneration and premature death in transgenic mice after proteolytic cleavage (reviewed in extracellular space Dieder Moechars, Kristin Lorent, Selkoe, 1994). Bart De Strooper, Ilse Dewachter and mediated by an uncharacterized This proteolysis step, Fred Van Leuven' is an intrinsic and essential proteinase named a-secretase, Experimental Genetics Group, Center for Human Genetics, secretion of APP and has become step in the mechanism of Campus Gasthuisberg, B-3000 Leuven, Belgium K.U.Leuven, of APP, both the focal point in the analysis of processing in the pathology 'Corresponding author fundamentally and for its involvement in the of Alzheimer's disease (AD). The senile plaques hallmark, A double mutation in the a-secretase site in the brain of AD patients that are a pathological PA4 region of mouse amyloid precursor protein (APP) contain mainly the ,A4 peptide (Glenner and Wong, 1984; amino acid peptide its secretion from COS cells, polarized MDCK Masters et al., 1985). This 39-43 reduced neurons. Expression of this represents part of the transmembrane and ectodomain of cells and rat primary using the neuron-specific APP (Goldgaber et al., 1987; Kang et al., 1987; Tanzi mutant in the brain of mice, et al., 1987). Since most patients with trisomy 21 develop elements of the mouse Thy-i gene promoter, resulted in AD-like senile plaques in their brain (Glenner and Wong, transgenic mice that became progressively hyperactive, 1984), this is to be attributed to a gene dosage effect of displayed seizures and died prematurely. In three the APP gene (Rumble et al., 1989). This indication for different transgenic lines the severity of the phenotype the narrow margin of regulation of expression of APP is was related directly to the expression levels of the corroborated further by the rare, early onset and presenile estimated by both mRNA and protein levels. transgene, familial AD cases that are inherited in an autosomal In addition, homozygous mice derived from each trans- dominant manner (Chartier-Harlin et al., 1991; Goate strain showed more severe symptoms which also genic et 1991; Murrell et al., 1991; Mullan et al., 1993). in life than in heterozygotes. The al., occurred earlier membrane-bound APP is cleaved not essentially dif- Newly synthesized observed symptoms were, however, intracellularly within the ,A4 peptide sequence by in ferent the different lines. Increased aggressiveness, the soluble 110-130 kDa acid and N-methyl-D- a-secretase, thereby generating disturbed responses to kainic and leaving cell-bound 10-12 kDa C-terminal in ectodomain aspartate, neophobia and deficiency exploratory transmembrane and cytoplasmatic domains (Selkoe, 1994). behavior were demonstrated in these mice. In the of by a-secretase, production In the absence cleavage brain, the observed neuropathological changes of the 13A4 is possible by the action of other peptide included necrosis, apoptosis and astrogliosis in the cellular named and uncharacterized proteinases, hippocampus, cortex and other areas. The data demon- P- Cellular studies in vitro with the clinical APP y-secretases. strate that incomplete or incorrect a-secretase pro- demonstrated that the a-secretase cleavage is mutants of APP results in severe neurotoxicity and that cessing in favor of ,- and y-cleavage: considerably more reduced this effect is in a dominant manner. expressed is from the Swedish APP mutant peptide produced precursor protein/apoptosis/neuro- Keywords: amyloid OA4 et al., 1992) while the APP717 mutants (swAPP) (Citron mice degeneration/transgenic more of the longer 42-43 amino acid yield relatively PA4 et The Swedish mutation peptides (Iwatsubo al., 1994). also affects intracellular routing and processing of APP MDCK cells et al., 1994; De Strooper in polarized (Lo Introduction Haass et et al., 1995a; al., 1995). The function of amyloid precursor protein (APP) Whereas the known clinical mutations in APP affect only is the inhibition of conclusively demonstrated to date and no natural mutations have the y-secretase sites, P- for the molecu- with specificity at the a-secretase site. Located most closely serine proteinases arginine been described inhibitor domain. lar forms of APP that contain the Kunitz the clinical mutations Ala692-Gly and Glu693->Gln, are the P1 The accumulated molecular evidence complicates downstream of the a-secretase residue, Lys687. situated APP located for from the gene mutation is clinically responsible analysis considerably: single The Glu693->Gln in some cases com- at least seven isoforms can be genera- recurrent on chromosome 21, hemorrhages (HCHWA-D) The alternative mRNA (Sandbrink et al., with AD et Ala692-*Gly ted by splicing bined (Levy al., 1990). mixed clinical of AD and references the isoforms in a picture 1994 and therein). Moreover, mutation, resulting et was shown as transmembrane that can be recurrent hemorrhages (Hendriks al., 1992), are synthesized proteins secretion of APP et N- and to cause reduced (Haass al., 1994). processed, O-glycosyla- differently proteolytically of the a-secretase as well as all to To examine the importance processing sulfated, varying ted, phosphorylated, introduced a double mutation in a end of we have This heterogeneous product step, previously degrees. yields very two basic residues that is secreted into the the by replacing kDa ectodomain ca-secretase region a 110-130 © Press Oxford University D.Moechars et al. Fig. 1. Secretion of APP/WT and APP/RK in COS cells, polarized MDCK cells and cultures of rat neurons. Cells primary hippocampal were metabolically labeled and APP was from cell extracts immunoprecipitated labeled and from the conditioned medium (lanes CELL) (lanes labeled SUP) and analyzed by SDS-PAGE. In MDCK cells, precursor APP and the mature N- and APP can be O-glycosylated (APPm) (APPp) clearly separated, while the APP75 1 is indicated endogenous by asterisks. with two acidic residues. The APP/RK mutant resulting Table I. Overview of all founder mice with the APP/RK mutant was secreted from COS cells at a rate that was <50% cDNA relative to APP wild-type et (APP/WT) (De Strooper al., 1993). We here Founder No. report that, also from MDCK Copy Offspring Expression polarized (%) the secretion of the APP/RK cells, mutant is markedly a/RK/B/1 3 + <5 reduced without the affecting exclusively basolateral secre- a/RK/B/2 + <5 tion In use of a viral pattern. addition, making infection a/RK/B/3 -- system, we demonstrate that rat neurons a/RK/F/4 + primary secrete 3 a/RK/F/5 2 + <5 less ectodomain from than considerably APP/RK from a/RK/F/6 1 + APP/WT. The functional of the APP/RK repercussions a/RK/F/7 + 2 5 mutant were in mice analyzed transgenic by expressing ae/RK/B/I 5 + <5 the APP/RK cDNA the elements of using neuron-specific ae/RK/B/2 3 + <5 the mouse ae/RK/F/3 3 Thy-i gene promoter et The + (Vidal al., 1990). ae/RK/F/4 4 mice resulting transgenic displayed a disturbed behavior, ae/RK/F/5 2 which deteriorated progressively with and resulted in age ae/RK/F/6 15 + <5 premature as soon as 2 death, occurring months after birth ae/RK/F/7 20 + 5 for individual of some ae/RK/F/8 homozygotes transgenic lines. 10 + ae/RK/F/9 The 1 + results of behavioral molecular, biochemical, and ae/RK/F/10 2 + histochemical tests and are and experiments presented t/RK/F/l -- discussed. t/RK/B/2 1 + 50 t/RK/F/3 1 + 5 t/RK/F/4 + 1 45 Results t/RK/F/5 1 + 30 t/RK/F/6 2 + 10 Expression of APPIRK in COS cells, polarized MDCK cells and neurons primary COS cells were transiently transfected with constructs the containing cDNA coding for APP/RK or APP/WT cultures of rat hippocampal neurons were infected with under the control of the early SV40 promoter recombinant (De Strooper Semliki forest virus carrying the APP/WT or et al., 1993). Secretion of APP/RK was ~50% relative the APP/RK mutant cDNA. The secretion of APP/RK was to APP/WT, confirming our previous observations (De again <50% relative to secretion of APP/WT (Figure 1). et Strooper al., 1993). MDCK cell lines were created that have either APP/ Generation of APPIRK transgenic mice and WT or APP/RK cDNA driven by the RSV promoter stably analysis of expression integrated in their genome (De Strooper et al., 1995a). All transgenic lines derived using APP/RK cDNA in Three independent cell lines of each were grown polarized different constructs are summarized in Table I. Transgenic on filters, and the apical and basolateral compartments founders are denoted by a code identifying the promoter were analyzed for APP secretion. In standard conditions, (t, a or ae, respectively for the mouse Thy-i gene promoter, secretion of mutant APP/RK is considerably decreased the human APP gene promoter and the latter with a viral relative to APP/WT and remains exclusively in the baso- enhancer element), the APP cDNA type or mutant (in the lateral compartment (Figure 1), a situation not different present work only APP/RK), the mouse strain (F or B, from APP/WT (De Strooper et al., 1995a). Primary respectively for FVB or C57B1) and the line number. The 1266 Mice transgenic for a-secretase APP mutant mRNA in the wild-type and t/RK strains, with the larger mRNA size due to the 3'-untranslated region of the Thy-i gene. Expression by the human APP gene promoter was low relative to endogenous APP. The mouse Thy-i gene promoter resulted in brain expression levels in hetero- zygotes that reached 50% of endogenous APP mRNA levels (Table I, Figure 2). In homozygotes, APP/RK mRNA in brain approached those of endogenous levels APP. in the t/RK lines Restriction of expression to brain was demonstrated by RT-PCR and Northern blotting. Endogenous Thy-i gene was expressed in brain and thymus (Figure 2). The thymus-specific elements of the mouse Thy-i gene are present in intron 3 and deleted in the construct used (Vidal et al., 1990). Western blotting of brain extracts with different poly- clonal and monoclonal antibodies detected consistently more of the 110-130 kDa APP present in the brain of transgenic APP/RK mice relative to non-transgenic control mice (Figure 3A). Individual variations between mice were apparent but, overall, less important than the differences between the different transgenic strains or between mice with a different genetic status: heterozygotes expressed APP controls and levels that are intermediate between 3E). Semi- homozygotes of the same strain (Figure quantitative comparison indicated that brain APP levels in the heterozygotes were ~75% (line t/RK/F/6) and 200% The highest level (t/RK/F/4 strain) higher than in controls. of was noted in the brain of homozygous expression t/RK/F/4 mice 3C) amounting to ~3-fold the (Figure APP levels. These protein data largely endogenous paralleled the expression at the mRNA level. In all the APP/RK lines a larger APP protein band was detected that is hardly visible in control brain extracts (Figure 3). This uncleaved band, representing mature, proteolytically APP/RK is being characterized and its metabolism examined extensively, not only in the cellular paradigms in used above, but also primary cortical and hippocampal neuronal cultures derived from the transgenic mice. Demonstration of reactive gliosis in the transgenic mice us to the levels of (see below) prompted compare glial acidic Western fibrillary protein (GFAP) by blotting. Some increase in GFAP was in the evident, especially animals, but was not nearly as prominent as homozygous the increase in APP Combined immuno- (Figure 3). precipitation and Western blotting failed to detect sig- nificant amounts of the peptide in brain extracts from PA4 not normal and APP/RK transgenic mice (results shown). Fig. 2. Analysis of mRNA expression. (A) Northem blotting of total in brain was The expression of other genes the analyzed brain RNA from a non-transgenic and from five t/RK heterozygous mRNA the APP-related at the protein or level, including and from transgenic mice. (B) Northern blotting of RNA from brain APLP1 and the E proteins APLP2, apolipoprotein gene mouse. seven different tissues of a homozygous t/RKlB/2 transgenic 4 kb detected A2MRILRP and A2MRAP A mouse Thy-i gene exon probe (1.4 ApaI fragment) and its receptors VLDLR, kb I mRNA and the 3.9 kb mRNA. RT-PCR the 1.85 Thy- APP/RK (C) et The APP/RK strains did (Lorent al., 1995). transgenic brain mRNA PCR for APP/RK mRNA on using primers specific differ in this from not appreciably respect age-matched or APP mRNA (lanes primers L921-L57) endogenous (lanes 2, 1, control mice. Positive control was RNA from COS cells primers L922-L57). APP/WT or APP/RK transiently expressing (see Figure 1). of APPIRK Phenotypic characterization transgenic mice concern and homo- experiments mainly heterozygous death. Mice from the C57B1 and CBA/ of founders t/RK/F/ Premature FVB, zygous offspring t/RK/B/2, t/RKIF/4, As more strains live >2 6 and ae/RK/F/7 J normally years. progressively (Table I). this was of the in the APP/RK mice died RT-PCR established expression transgene prematurely, phenomenon a of 64 in four monitored in in four out of 10 ae/RK population heterozygous transgenic brain transgenic strains, relative to a of 30 in five out of six t/RK lines mice of the first out of six a/RK lines and generation, group littermates. In the first revealed 3.5-3.9 kb Northern genotyped, non-transgenic year, (Table I, Figure 2). blotting 1267 D.Moechars et al. only 7% of non-transgenic mice while 66 mild seizures of died, 44, and transfer varying severity and, upon 69% of transgenic mice died in lines into clean t/RK/F/6, t/RK/F/4 cages, posture and Sex freezing neophobia. and t/RKJB/2, respectively (Figure 4). The first death difference was not Seizures in consisted of facial apparent. lines tIRKIB/2, t/RK/F/4 and t/RK/F/6 occurred at mild fore-limb clonus and 30, 78 movements, mild whole body and 150 days of clonus age, respectively. This correlated from to 10 s. with lasting S Transgenic mice older than the higher 6 APP/RK expression level in lines months in severe t/RKlB/2 and showed, addition, tonic-clonic seizures, t/RK/F/4 relative to t/RK/F/6 30-60 (Table I). lasting s. These mice and stretched gasped all four All animals that died were limbs with spontaneously weighed and their tail followed upright, by severe whole necropsied, but no malformations of the vital clonic organs nor seizures. mice were body Afterwards, for lethargic major internal bleeding were observed. Premature variable of death periods time. Seizures have been observed in was not preceded by overt of deteriorated health >60% of all signs or mice in these two lines. transgenic Seizures by marked loss of weight. were absent or rare in mice of line t/RK/B/2 very (Table which were I), backcrossed into the C57Black background In Spontaneous behavior lines t/RK/F4 and t/RK/F/6, were at (mice analyzed least 87.5% of this background). heterozygotes of <3 months and of <2 homozygotes This strain is known to be more resistant to seizures and months behaved normally. showed Subsequently, they to drugs inducing seizures (Ganesan et 1995 and al., increased agitation, bouts of wild running and, among references therein). Spontaneous seizures were never more males, frequent and fierce fighting. Many displayed observed in normal mice of the FVB strain in our which are under colony, kept identical conditions to the transgenic mice. Behavioral In a new testing. environment, normal mice move immediately towards the wall of the cage and the walls proceed along (thigmotaxis) (Walsh and Ambulation as an Cummins, 1976). indicator of motor was activity scored of by parameters latency, i.e. the time elapsed before the beginning of thigmotaxis and by the number of corners crossed during the 30 s interval. In lines t/RK/F/4 and tIRK/F/6, respectively, 54 and 47% of the mice exhibited a latency of >5 as to 5% s, opposed of control FVB mice. Younger transgenic mice behaved like age-matched littermates: at 4-6 weeks, <10% of mice of both lines transgenic displayed a latency of >5 s. Less than 2% of FVB controls showed posture freezing in a new environment, as opposed to 30% of t/RK/F/4 and 23% of t/RK/F/6 mice (3-5 months old). Transgenic mice of all ages also showed a markedly increased respiratory rate, while some mice went into seizures, indicating the severe stress. In a group of 35 FVB control mice, the frequency of corner crossing was Gaussian distributed, centered around a mean of seven crosses (Figure 5). This distribution was Fig. 3. Western blotting of APP and GFAP in brain of control and transgenic mice. (A) FVB control mice (lanes 1-3) and homozygous t/RK/F/4 transgenic mice (lanes 4-6) immunoblotted with mAb 22C1 1 (Hilbich et al., 1993) on a 6% homogeneous SDS-PAGE gel. (B) The same extracts separated on a 4-20% linear SDS-PAGE gradient gel, immunoblotted with a polyclonal antibody against GFAP. (C) Different amounts of brain extract of FVB control (extract 1 from A) and homozygous t/RK/F/4 (sample 4 in A) on a 6% gel, stained with mAb 22C11. Lane 1: 6 lane 2: 2 lane 3: 0.6 ,l, lane 4: 0.2 p1, gA, gl, lane 5: 0.2 lane 6: 0.6 pl, lane 7: 2 Note the equivalent staining gl, g1. intensity of lanes 3 and 5 (volume extract ratio of 3) and the stronger staining of the larger extra band of APP in lanes 5-7. (D) Brain extracts of different control and heterozygous transgenic mice, separated on a 4-20% gradient gel and stained with mAb 22C Lane 1: FVB control, lane 2: Fl(C57BlXCBA/J) control, lane 3: heterozygote t/RK/B/2, lane 4: heterozygote t/RK/F/4, lane 5: heterozygote t/RK/B/6, lane 6: heterozygote of a mouse, t/APP770/F/1 transgenic for mouse APP770 cDNA driven by the mouse Thy-l gene promoter. (E) Brain extracts of different heterozygous and homozygous transgenic mice separated on a 4-20% gradient gel and stained with mAb 22C11. Lanes 1 and 2: heterozygous and homozygous t/RK/F/4, lanes 3 and 4: heterozygous and homozygous t/RK/F/6, lanes 5 and 6: heterozygous and homozygous t/RK/B/2, lane 7: FVB control. 1268 transgenic for a-secretase APP mutant Mice 80. 25 - a) 20 - M 70 15 - I'2 8) 5- .I.I.I.I. I. AL X <~~~~~~~~~~5. -zz---x 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 =30 control n .> 30 13 -4-I-t/RK/B/2 n =10 -X-tURKIFI3 CO 20 -?a-I n 18 VRK/F/4 10 -X-tVRK/F/6 n 18 -C-O-tRKIF/4 Ho n =22 0 I1 1 1 1 1 1 1 I 0 0 0 0 0 0 0 0 0 0 0 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 (D) 0 03) (a LO co 0 C _N c' cn c' c Age (days) Fig. 4. Premature death of control FVB, heterozygous and 25- homozygous APP/RK mice. The death rate in four t/RK lines that were heterozygous for the APP/RK transgene compared with non- 15- transgenic littermates over a 1 year period. Homozygous offspring of the t/RK/F/4 line were followed for only 105 days. 4 5 6 7 8 9 10 11 12 13 14 0 1 2 3 15, 0 1 2 3 4 S 6 7 a 9 10 11 12 13 14 15 similar for young or old FVB mice (2-3 or 9-12 months Corners crossed old). Heterozygous and homozygous mice of lines t/RK/ t/RK/F/6 (6-20 weeks old) displayed significantly F/4 and ambulation. Strikingly, the frequency of corner reduced crossings was skewed towards low values, separating two one with low explorative behavior and subpopulations, with only a slight reduction. Afterwards, the tested another 150 animals did not behave differently from untested transgenic 5 120-- mice. It was also evident that premature death was not ~'0 confined to either subpopulation of the tested transgenic mice: two out of five t/RKJF/4 mice and two out of M 30 -- 11 t/RK/F/6 mice (from those tested in Figure 5) that died ni during the 6 months subsequent to the open field test, Test I Test belonged to the low explorative groups. This proved that the reduced exploratory behavior was not a direct sign of death of the transgenic animals. imminent, premature aggressive behavior in the t/RK/F/4 The augmented 30 () 25 offspring was measured under controlled and t/RK/F/6 o 20 Single male FVB control and transgenic mice, conditions. for the transgene, were housed isolated for homozygous <10 4 before being confronted in their cage 5 weeks ('resident') with a male FVB 'intruder' that had been reared in a was scored by the latency of the group. Aggression I Test 2 Test first attack and by the number of attacks during 3 min 5. Behavioral testing. (A) In open field testing, the number of (Saudou et al., 1994). APP/RK resident males Fig. observation crossed during a 30 s interval was measured. FVB control corners attacked sooner and more often than non-transgenic FVB mice panel, n = 35), t/RK/F/4 mice (middle panel, n = 30) (upper resident males. Transgenic males from strain t/RK/F/4 n = (B) In the aggression test, mice (lower panel, 50). and t/RK/F/6 attacked sooner and more often than t/RK/F/6 residents were FVB bars, n = 4), homozygotes of lines resident mice (black the attacks were judged to be fiercer and n = and t/RK/F/6 (striped bars, n = 5). 5), and t/RK/F/4 (open bars, 6) (Figure attack panel) and the number of attacks by the Latency of (upper intense. When the same animals were retested, more in a 3 min observation period. Data are resident (lower panel) was increased (Figure 5). aggressiveness with standard errors. Statistical analysis mean values calculated between and control mice to be showed the difference transgenic to glutamate analogs Reactivity (P<0.02). significant The system is the major excitatory system glutamatergic in the generation of seizures. Its activity was involved for KA induction of seizures and curves Dose-response tested in lines t/RKIF/4 and t/RK/F/6 with two agonists and mice in C57B1 FI(C57BlXCBA/J) mortality FVB, of different glutamate receptors, i.e. kainic acid (KA) and strain differences being more obvious (FVB demonstrated (NMDA). N-methyl-D-aspartate 1269 D.Moechars et al. hilus of the dentate gyrus and dispersed throughout the 80 - granular layer of the dentate gyrus. In the most affected 60 - animals, neurons were in the degenerating present thalamus 40 - (Figure 7) and, to a lesser extent, in the posterior cingulate 20- cortex, in the caudate putamen, the and hypothalamus the CAI and CA2 region of the 0 hippocampal pyramidal layer. c 32 Typical also were the nuclei of glial cells in the 24 40 corpus callosum and in the fimbria which are normally ellipsoid, Dose of KA (mg/kg) 0 but appeared rounded and more densely stained in APP/ RK mice (Figure 7). The most severe signs of neuro- degeneration were encountered in four t/RK/B/2 and in two t/RK/F/4 mice. The age of the animals was not the determining factor, since these mice were aged between 2.5 and 9 months. The abnormal neurons in the dentate gyrus were reminis- cent of pyknotic or apoptotic cells: condensation of the chromatin results in compact and heavily stained nuclei 40 50 60 80 100 with blebbing of the nuclear membrane. Apoptotic cells Dose of NMDA (mg/kg) have been demonstrated by immunochemical methods in AD brain (Su et al., 1994) and in transgenic mice (LaFerla Fig. 6. Effects of glutamate analogs kainic acid and NMDA. Upper of KA to FVB mice et al., 1994). In situ detection of apoptotic cells in brain panel: mortality resulting from i.p. administration = mice (black bars, n 4, 9 and 6), homozygous t/RK/F/6 bars (open slices by terminal transferase incorporation of digoxigenin- = mice n = n 2, 8 and 4) and homozygous t/RK/F/4 (striped bars, 2, demonstrated dUTP and peroxidase-based detection, con- from of 9 and 4). Lower panel: mortality resulting i.p. injection siderable numbers of apoptotic cells in t/RK/B/2, t/RK/ in n = 4 and NMDA FVB mice (black bars, 4, 6, 7, 2), homozygous mice In mice of line t/RK/ F/4 and ae/RK/F/7 (Figure 7). n = 4 and and t/RK/F/6 mice (open bars, 2, 4, 4, 3) homozygous mice = 6 and of line B/2 and some of line the of the t/RK/F/4 (n 4, 6, 5, 2). Homozygous offspring t/RK/F/4, granular layer were and and died when t/RK/F/4 resistant, displayed symptoms only dentate contained while in gyrus many apoptotic neurons, receiving higher doses (see text). ae/RK/F/7 mice apoptotic neurons were in the present amygdala (Figure 7). sensitive than but KA Reactive gliosis or astrogliosis was demonstrated by C57Bl mice), consistently elicited, dose clonic seizures which were rated in five immunostaining for GFAP, and proved positive even in dependently, stages: immobilization, head forelimb younger transgenic mice and in mice from line t/RK/B/2 nodding, clonus, rearing and rearing with falling (Racine, 1972). All APP/ that did not exhibit seizures. An intense astrocytic reaction KA was in RK mice were more sensitive to than were control observed the cortex, the amygdala and the hippo- KA seizures of mice (Figure 6). at 40 mg/kg triggered campus; astrocytes with enlarged cell bodies and thicker 5 while stained for Brains stage in all t/RK/F/4 and t/RK/F/6 mice, only processes intensely GFAP (Figure 7). of FVB control mice showed 4 seizures and of mice from lines and 66% stage transgenic t/RKIF/4 t/RK/F/6 were 30% 5. the all t/RK/F/4 mice stained for the with five stage Following seizures, immunochemically peptide PA4 died, while 35% of t/RK/F/6 mice and 50% of control different antibodies. some was evident in Although staining FVB mice survived (Figure 6). certain neurons, the reaction was weak not (results shown). The for NMDA in FVB control mice was LD50 ~50 mg/kg body weight (Figure 6). The APP/RK trans- Discussion mice genic demonstrated a markedly diminished sensitivity for NMDA. mice even resisted a Homozygous t/RK/F/4 The functional importance of correct a-secretase pro- dose of 100 which was of APP was mg/kg, invariably lethal for control cessing examined by creating a mutant APP FVB as well as for all t/RK/F/6 mice that was transgenic (Figure partially defective in secretion from three different 6). In t/RK/F/4 mice this dose of NMDA failed to cell COS high types: cells, MDCK cells and primary neurons. elicit the typical NMDA symptoms evident in control Expression of this a-secretase mutant in mouse brain FVB mice (Leander et al., 1988). These symptoms were resulted in a progressive disorganization of the central evident at 150 a dose mg/kg, that was never tested in nervous system, resulting in behavioral disturbances, control or t/RK/F/6 mice and which killed all t/RK/F/4 seizures, differential reaction to glutamate analogs, prema- transgenic mice in 20 min. ture death and morphological neuronal changes in the brain. The histological analysis demonstrated conclusively Histological analysis of the brain of transgenic that the brain of transgenic t/RK mice contained many APPIRK mice abnormal neurons in the dentate and in the thalamus gyrus In the brain of 24 out of 25 transgenic mice of lines and, in older mice, also in the cortex and the pyramidal t/RK/B/2 (four mice), t/RK/F/4 (15 mice), t/RK/F/6 (three layer of the hippocampus. The symptoms of deranged mice) and ae/RK/F/7 (three mice) that died during observa- behavior, the occurrence of seizures and the premature tions, typical abnormal neurons were evident: many death are a consequence of the documented neuronal neurons were vacuolized or contained clear cytoplasm, damage. This conclusion is derived from the accumulated surrounding compacted and densely stained nuclei (Figure data in presented the Results section and discussed below. 7). These aberrant neurons were most conspicuously Expression of the transgene was restricted exclusively located on the boundary of the granular layer and the to the brain, as anticipated from the characteristics of the 1270 Mice transgenic for a-secretase APP mutant A B * 0 S *1'6* 9% ~~~~~~~~~~~4 0409 ~~~~~0 4 40* S. ... ,s:t ;, 4% 94 h.~ t 'IA *ee 01.~~~~ 49~~~~~~~ E F J H a0I 4F .S . i i':l Oeb * e St ",W. _.,,. ' X # .! k * v , 2 * # s . * it 0 41k... -1*it * .> \ .. ... t.00 ;. 4 ..s s; S 0 A.tt 0 * oe K L :thi ~ ~ : ~ ~ ?. p S :V .A- :lk S, ,A rk A; * 7. and GFAP of mouse brain. Fig. Histology, apoptosis immunochemistry of sections of the brain of FVB (A-J) Hematoxylin-eosin staining sagittal and APP/RK mice. show the of FVB Representative photomicrographs granular the dentate from control mouse t/RK/B/2 mouse layer gyrus (A), (B), t/RKIF/4 mouse and t/RK/F/6 mouse of (C) (D); the thalamus from control FVB mouse and t/RK/F/6 mouse of the callosum (E) (F); corpus from control FVB mouse and t/RK/F/6 mouse of the CAl from control FVB mouse and t/RK/B/2 mouse (G) (H); hippocampal pyramidal layer (I) (J) (X60). (K-N) In situ apoptosis in the brain of APP/RK mice detected on sections in the dentate from control FVB transgenic sagittal gyrus (K), t/RK/B/2 and t/RK/F/4 mice and in the of an mouse GFAP in the cortex of control (L) (M) amygdala ae/RK/F/7 (N) (x 100). (O-S) immunostaining a of the FVB (0) and t/RK/F/4 mice in the dentate of FVB and mice is (P) (x 10), gyrus (Q) t/RK/F/6 (R) (X20). (S) higher magnification (X40) bottom right hand corner of (R). the mouse used in the mRNA and level. the elements of at protein Thy-i gene promoter Although homozygous the most severe and the constructs et In the four of strain t/RK/F/4 injected (Vidal offspring display al., 1990). mild as in mouse strains and even independent transgenic spectacular phenotype, analyzed overexpression, resulted in the of the was related of line that described, severity homozygotes t/RKIF/6, symptoms phenotype same. In not to the of the in the brain both are the directly expression essentially addition, although presented transgene 1271 D.Moechars et al. processed a-secretase. here, a similar phenotype is becoming evident in by Obviously, this is unique since transgenic the APP/RK mice that express the same APP/RK mutant but mutant is artificial and it is impossible to cDNA, how the double under the control of the human APP gene promoter predict mutation at the C-terminal end (the a/RK and The in these strains of the secreted APP/RK ae/RK lines). symptoms are ectodomain would disturb an milder and at a later which is consistent function. develop age, with (unknown) Alternatively, the slower processing with a-secretase the lower expression level attained this by should favor [-cleavage of promoter APP/RK, in (unpublished results). Further, we have generated trans- resulting secretion of [3-cleaved APP/RK, a situation to genic mice in which the mouse be in Thy-i gene promoter envisaged patients carrying the Swedish APP muta- elements are used to drive of the tion et expression mouse (Citron al., 1992). APP770 cDNA or of the C-terminal domain These considerations should take (Figure 3) of into account the APP fused to its signal These intracellular fate of peptide (not shown). unprocessed APP/RK and of cell- mice show none of the bound the transgenic phenotypic character- fragments containing [A4 peptide resulting istics of the APP/RK mice. We conclude that the observed from [-cleaved APP. An interesting collateral is provided phenotype is specific and caused the over- by the recent report of transgenic mice directly by that overexpress expression of the proven a-secretase APP/RK mutant in the [BA4 peptide intracellularly et (LaFerla al., 1995). The the brain. many phenotypic traits that are similar to those observed Some of the of the APP/RK in the APP/RK mice described here invite phenotypic parameters speculation mice are reminiscent of i.e. dis- about a common cause. Possibly, the cell-bound AD, neurodegeneration, APP/RK turbances in behavior and premature death. C-terminal fragments can lead to intracellular Preliminary [A4 peptide, tests indicated that the short-term of the creating a situation that is not essentially different from learning ability the is one mice not affected, but more extensive and careful in which the [A4 peptide is produced intracellularly as testing of and is in learning memory capabilities progress. The instructed by the transgene (LaFerla et al., 1995). This occurrence of seizures in our generalized transgenic mice reasoning implies that intracellularly produced [3A4 pep- was but literature data that these are tide could unexpected, prove cause neurodegeneration and, eventually, could frequently observed in AD even as a be more neurotoxic than patients, prominent extracellularly produced [A4 feature in familial chromosome 14-linked while less peptide. This AD, would explain why transgenic mice with a prevalent in with the APP717 mutations massive 10-fold patients (Mullan overexpression in brain of a clinical APP et al., 1993; Lampe et It was debated whether mutant protein on the one al., 1994). displayed, hand, massive seizures constitute an characteristic of AD or deposits of [-amyloid in authentic integral extracellular amyloid are a secondary phenomenon. The loss of neurons and plaques but, on the other, produced a only very mild important alterations of the architecture of phenotype (Games et al., 1995). more hippocampal Clearly, extensive and neocortical areas provides a suitable substrate for the comparative and longitudinal studies are on required the generation of seizures, while seizures could enhance the different transgenic mouse strains. of the disease. progression An observation that could be related to directly the Glutamate has been in mechanisms involved as excitotoxicity implicated the discussed is the production of neurodegeneration and in neuronal cell in or [3A4 process death amyloid peptide in mice. Little or no [BA4 peptide- AD. In the tests with the glutamate the observed are in analogs, containing peptides produced COS cells, MDCK refractiveness to NMDA of the APP/RK mice cells or neurons after transfection transgenic with mouse APP as could be 'NMDA indicative of opposed to human APP et receptor hypofunction' (De Strooper al., 1995b and recently as a proposed novel type of excitotoxicity unpublished results). On the other hand, intracellular (Rothman and In Olney, 1995). addition, since NMDA amyloid-containing fragments are formed readily from receptor activity is to be mouse thought linked intimately to the APP695 after 'humanization' by changing only central mechanisms of and three amino memory building learning acid residues in the [BA4 peptide region of (Malenka, 1994; Collingridge and Bliss, we believe murine 1995), APP695 (De Strooper et al., 1995b). The fact that that the APP/RK mice will allow further studies in we used mouse APP this to generate the mutant APP/RK direction. The difference in reactivity in this mice respect, transgenic contributed to or highlighted precisely caused by the APP/RK transgene in the different genetic that aspect that otherwise might have been blurred, i.e. backgrounds (FVB or C57Black, respectively in the t/RK/ that the normal and pathological physiology of APP in F/4 or t/RK/F/6 and t/RKIB/2 transgenic lines), is most brain includes more than production of the [A4 peptide. interesting: the major and so far only difference is the In this respect, it is noteworthy that APP-deficient mice almost complete absence of seizures in line t/RK/B2 while, demonstrated several phenotypic characteristics in com- by the other parameters and criteria of molecular and mon with the APP/RK transgenic mice (Zheng et al., histochemical analysis, this line is not notably different 1995 and unpublished results). Eventually, not only the from the FVB transgenic lines. The C57Black mouse production of peptide but also the absence or suppres- PA4 strain is known to be more resistant to seizures and to drugs sion of an as yet unknown function of APP will explain inducing seizures (Ganesan et al., 1995 and references this paradox. therein). This clear genetic component in the phenotype is another strong argument that the seizures are the consequence and not the cause of the neurodegeneration. Materials and methods To explain the effect of the APP/RK mutation molecu- Construction of the APP695/RK minigenes larly, two possibilities need to be considered. Given the The mouse APP/RK cDNA (De Strooper et al., 1993) was cloned in an location of the double mutation, the secreted APP/RK adapted site in the vector (gift SfiI pTSCaxI from H.Van der Putten, ectodomain will contain a mutated C-terminal end when Basel). This pUC18-based vector contained an 8.1 kb EcoRI fragment 1272 Mice transgenic for a-secretase APP mutant comprising the mouse Thv-1.2 gene (Vidal et al., 1990). A 1.5 kb BanI- (Ciba-Geigy, Basel), C.Dotti (EMBL, Heidelberg), B.Greenberg XhoI fragment (located in exon 2 and exon 4, respectively) was replaced (Cephalon, West Chester, PA), M.Crauwels (Leuven) and S.Dunnett by the APP/RK cDNA. Alternatively, the mouse cDNA was (Cambridge). APP/RK Expert technical assistance was provided by M.Gilis, cloned downstream of the human APP gene promoter with a ,B-globin C.Kuiperi, K.Craessaerts, S.Torrekens and L.Stas. This investigation intron interspersed. A duplex polyoma virus enhancer element eventually was supported by the 'Fonds voor Geneeskundig Wetenschappelijk preceded the human APP promoter. Onderzoek' (FGWO-NFWO), the Interuniversity network for Funda- mental Research (IUAP, 1991-1996) and the Action Program for Microinjection and analysis of founders Biotechnology of the Flemish government (VLAB, ETC-008). B.D.S. is Linearized constructs were purified and microinjected into 1.5-day-old a postdoctoral fellow of the National Fonds voor Wetenschappelijk pre-nuclear embryos taken from superovulated FVB or Fl (C57BIX Onderzoek (NFWO) and thanks the Sandoz Foundation and the Divry CBA/J) females. The injected oocytes were allowed overnight to reach Research Fund for support. D.M. is the recipient of an IVT/IWONL the two-cell stage and transferred into pseudopregnant fosters, Fl scholarship. (C75BIXCBA/J). Tail biopsies were taken 3 weeks after birth at weaning for isolation of genomic DNA and Southern blotting, as described References (Umans et al., 1994). Chartier-Harlin,M.C. et al. (1991) Ea-rly-onset Alzheimer's disease Analysis of RNA and protein caused by mutations at codon 717 of the precursor protein 0-amyloid Total RNA from mouse tissues was isolated and analyzed by Northern gene. Nature, 353, 844-846. blotting as described (Lorent et al., 1995). RT-PCR was carried out with Citron,M., Oltersdorf,T., Haass,C., McConlogue,L., Hung,A.Y., sense primers containing either the wild-type APP or the mutated APP/ Seubert,P., Vigo-Pelfrey,C., Lieberburg,I. and Selkoe,D.J. (1992) RK region (De Strooper et al., 1993): primer L921, 5'ttgaagtcCGCcat- Mutation of the ,B-amyloid precursor protein in familial Alzheimer's caaAaA3' or primer L922, 5'ttgaagtcGATcatcaaGaG3' (the five nucleo- Disease increases :-protein production. Nature, 360, 672-674. tide differences are capitalized). The antisense primer was L57: 5' Collingridge,G. and Bliss,T. (1995) Memories of NMDA receptors and ccgatgggtagtgaagcaatggtt 3'. RT-PCR was done with touch-down LTP. Trends Neurosci., 18, 54-56. temperature programming: four series of three cycles each (1 min De Strooper,B., Umans,L., Van Leuven,F. and Van Den Berghe,H. (1993) denaturation at 92°C, 2 min annealing at 68, 67, 66 and 1 65°C and min Study of the synthesis and secretion of normal and artificial mutants extension at 72°C) followed by 35 cycles 92°C, 2 (I min min 64°C, of murine amyloid precursor protein (APP): cleavage of APP occurs 1 min 72°C). For Western blotting, brain tissue was homogenized in in a late compartment of the default secretion pathway. J. Cell Biol., vol of buffer (50 mM Tris, pH 8.0, 150 mM NaCl, 1% deoxycholate, 121, 295-304. 1% Triton X-100, 0.1% SDS) with proteinase inhibitors (25 mM EDTA, De Strooper,B., Craessaerts,K., Dewachter,I., Moechars,D., ,ug/ml pepstatin, 1000 U/ml aprotinin) in a motor-driven mechanical Greenberg,B., Van Leuven,F. and Van Den Berghe,H. (1995a) homogenizer. After centrifugation (12 000 g, 20 the clear min), super- Missorting of amyloid precursor protein in MDCK cells. J. Biol. natant was made 1% in SDS and 2-mercaptoethanol, boiled for 5 min Chem., 270, 4058-4065. and analyzed on Tris-glycine-buffered polyacrylamide gels (6% or De Strooper,B., Simons,M., Multhaup,G., Van Leuven,F., Beyreuther,K. 4-20%). and Dotti,C. (1995b) Production of intracellular amyloid containing fragments in hippocampal neurons expressing mutants of human Brain histology and immunostaining amyloid precursor protein causing Alzheimer's disease and protection Brain tissue was fixed ovemight in 4% paraformaldehyde, dehydrated against amyloidogenesis by subtle amino acid substitutions in the and paraffin embedded for sectioning (5 ,um). For in situ apoptosis rodent sequence. EMBO J., 14, 4932-4938. detection, sections were mounted on silanized slides and treated for Games,D. et al. (1995) Alzheimer-type neuropathology in transgenic terminal transferase incorporation of digoxigenin-dUTP and immuno- mice overexpressing V717F ,B-amyloid precursor protein. Nature, 373, peroxidase detection and staining (Apoptag kit, Oncor). For GFAP 523-527. immunostaining, 5 paraffin sections were dewaxed, rehydrated and Ganesan,L., Kofuji,P., Wang,J., Fernando,J., Liu,Z., Mathura,J. and gm treated with hydrogen peroxide to eliminate endogenous peroxidase Burt,D. (1995) GABA receptor ,B2 and ,B3 subunits: comparison in PI, activity. After blocking with normal goat serum (5% in 50 mM Tris, DBA/2J and mice. C57B1/6J Biochim. Biophys. Acta, 1261, 134-142. pH 7.4, 150 mM NaCl, 0.1% Triton X-100), sections were incubated Glenner,G.G. and Wong,C.W. Alzheimer's disease and Down's (1984) overnight at 4°C with rabbit anti-cow GFAP antibody (Dakopatts, syndrome: sharing of a unique cerebrovascular fibril amyloid protein. Denmark) diluted 1:1000, followed by peroxidase-conjugated goat anti- Biochem. Biophys. Res. Commun., 122, 1131-1135. rabbit antibody (dilution 1: 100). The reaction was developed with Goate,A. et al. (1991) Segregation of a missense mutation in the amyloid diaminobenzidine and hydrogen peroxide. precursor protein gene with familial Alzheimer's disease. Nature, 349, 704-706. Open field test Goldgaber,D., Lerman,M., McBride,O., Saffiot,U. and Gadjusek,D. All mice used were kept under a light cycle of 12 with h, food and (1987) Characterization and chromosomal localization of a cDNA water ad libitum. The corner index test is a variant open field test to of encoding brain an'yloid Alzheimer's disease. Science, 235, 877-880. measure activity and ambulation. The mice were placed in the middle Haass,C., Hung,A.Y., Selkoe,D.J. and Mutations Teplow,D.B. (1994) of a clean, unused opaque cage (20 cm square) and allowed to move associated with a locus for familial Alzheimer's disease result in under dim, indirect light for 30 s. Between tests, cages were cleaned. alternative processing of J. Biol. amyloid protein precursor. Chem., Controls were age matched and subjected at random to the tests. Results 269, 1-8. were scored independently by two observers, one of which was unaware Haass,C., Koo,E., Capell,A., Teplow,D. and Polarized Selkoe,D.J. (1995) of the genetic status of the mice. The mice were tested twice with a 3 sorting of 3-amyloid precursor protein and its in proteolytic products day interval. MDCK cells is regulated by two independent signals. J. Cell Biol., 128, 537-547. Aggression test Hendriks,L. et al. (1992) Presenile dementia and cerebral haemorrhage FVB or transgenic male 'residents' (tested at 10-12 weeks were old) linked to a of the mutation at codon 692 f-amyloid precursor protein housed alone for 4 weeks in normal cages in the same room under the gene. Nature 218-222. Genet., 1, same conditions as other mice. Bedding was changed once a week Hilbich,C., Monning,U., and Grund,C., Masters,C. Beyreuther,K. (1993) without handling the first the animal, with the last change 6 days before Amyloid-like of the of properties peptides flanking epitope amyloid in of test. 'Intruders' were male FVB mice (8 weeks old) kept groups precursor protein specific monoclonal 22C 11. J. Biol. antibody Chem., eight in larger After the intruder was in the resident cages. placed cage, 268, 26571-26577. attack number of attacks were recorded 3 min. An latency and during and Iwatsubo,T., Odaka,A., Suzuki,N., Mizusawa,H., Ihara,I. Nukina,N. attack when the resident bit the intruder. One week later was scored only of A beta and A beta 40 in senile (1994) Visualization 42(43) plaques the same resident mice were tested a second time with different intruders. that an with A beta monoclonals: evidence end-specific initially 45-53 deposited species is A beta 42(43) Neuron, 13, Kang,J., Lemaire,H.G., Salbaum,J.M., Masters,C.L., Unterbeck,A., Acknowledgements and Grzeschik,K.H., Muller-Hill,B. Multhaup,G., Beyreuther,K. Alzheimer's disease A4 The contribution or materials and the criticism of The of of expertise positive (1987) amyloid protein precursor 733-736. the following scientists is H.Van der Putten resembles a cell-surface 325, gratefully acknowledged: Nature, receptor. 1273 D.Moechars et al. and LaFerla,F., Tinkle,B., Bieberich,C., Haudenschild,C. Jay,G. (1995) The Alzheimer's peptide induces neurodegeneration and apoptotic AP cell death in transgenic mice. Nature Genet., 9, 21-29. of Lampe,T. et al. (1994) Phenotype chromosome 14-linked familial in Ann. Alzheimer's disease a large kindred. Neurol., 36, 368-320. Leander,J., Lawson,R., Ornstein,P. and Zimmerman,D. (1988) N-Methyl- in D-aspartic acid-induced lethality mice: selective antagonism by Brain phenylcyclidine-like drugs. Res., 448, 115-120. Levy,E., Carman,M.D., Femandez-Madrid,I.J., Power,M.D., Van Lieberburg,I., Duynen,S.G., Bots,G.T.A.M., Luyendijk,W. and Frangione,B. (1990) Mutation of the Alzheimer's disease amyloid in gene hereditary cerebral haemorrhage, Dutch type. Science, 248, 1124-1126. Lo,A., Haass,C., Wagner,S., Teplow,D. and Sisodia,S. (1994) Metabolism of the 'Swedish' variant in amyloid precursor protein Madin-Darby canine cells. 30966-30973. kidney J. Biol. Chem., 269, De Van Lorent,K., Overbergh,L., Moechars,D., Strooper,B., Leuven,F. and Van den in mouse and in Berghe,H. (1995) Expression embryos adult mouse brain of three members of the amyloid precursor protein of the related family, alpha-2-macroglobulin receptor/LDL receptor and of its protein ligands apolipoprotein E, lipoprotein lipase, alpha- and the 40 kDa associated 2-macroglobulin receptor protein. 1009-1025. Neuroscience, 65, Malenka,R. (1994) Synaptic plasticity in the hippocampus: LTP and LTD. 535-538. Cell, 78, and Masters,C., Simms,G., Weinman,N., Multhaup,G., McDonald,B. Beyreuther,K. (1985) Amyloid plaque core protein in Alzheimer disease and Down Proc. Natl Acad. Sci. 4245- syndrome. USA, 82, Mullan,M. et al. (1993) Clinical comparison of Alzheimer's disease in Ile in pedigrees with the codon 717 Val mutation the amyloid precursor protein gene. Neurobiol. Aging, 14, 407-419. Murrell,J., Farlow,M., Ghetti,B. and Benson,M.D. (1991) A mutation in the associated with amyloid precursor protein hereditary Alzheimer's Disease. 97-99. Science, 254, Modification of seizure Racine,R. (1972) activity by electrical stimulation: II. Motor seizure. Clin. Electroencephalogr Neurophysiol., 281-294. 32, and and the NMDA Rothman,S. Olney,J. (1995) Excitotoxicity receptor still lethal after Trends 57-58. eight years. Neurosci., 18, Rumble,B., Retallack,R., Hilbich,C., Simms,G., Multhaup,G., Martins,R., Hockey,A., Montgomey,P., Beyreuther,K. and Masters,C.L. (1989) A4 Amyloid protein and its precursors in Down's syndrome and Alzheimer's disease. New Engl. J. Med., 320, 1446-1452. and Sandbrink,R., Masters,C. Beyreuther,K. (1994) protein PA4-amyloid mRNA isoforms without exon 15 are precursor ubiquitously expressed in rat tissues but not in neurons. J. including brain, Biol. Chem., 269, 1510-1517. Ait Saudou,F., Amara,D., Dierich,A., LeMeur,M., Ramboz,S., Segu,L., and Enhanced behavior in mice Buhot,M. Hen,R. (1994) agressive lacking receptor. Science, 265, 1875-1878. 5-HTIB Cell of the and Selkoe,D. (1994) biology amyloid ,B-protein precursor the mechanism of Alzheimer's disease. Annu. Rev. Cell Biol., 10, 373-403. and Su,J., Anderson,A., Cummings,B. Cotman,C. (1994) Immunohisto- chemical evidence for in apoptosis Alzheimer's disease. NeuroReport, 2529-2533. 5, Tanzi,R.E. et al. (1987) Amyloid n-protein gene: cDNA, mRNA distribution and genetic linkage near the Alzheimer locus. Science, 235, 880-884. Umans,L., Serneels,L., Hilliker,C., Stas,L., Overbergh,L., De Strooper,B., Van and Van Leuven,F. den Berghe,H. (1994) Molecular cloning of the mouse for gene coding alpha-2-macroglobulin and targeting of the gene in embryonic stem cells. Genomics, 22, 519-529. Vidal,M., Morris,R., Grosveld,F. and Spanopoulou,E. (1990) Tissue- control specific elements of the Thy-] gene. EMBO J., 9, 833-840. Walsh,R. and Cummins,R. (1976) The open-field test: a critical review. 482-502. Psychol. Bull., 83, Zheng,H. et al. (1995) f-Amyloid precursor protein-deficient mice show reactive gliosis and decreased locomotor activity. Cell, 81, 525-531. Received on September 20, 1995; revised on November 7, 1995

Journal

The EMBO Journal – Springer Journals

Published: Mar 1, 1996

There are no references for this article.