Skip to main content
NCBI home page
Molecular Therapy logoLink to Molecular Therapy
. 2009 Jun 16;17(9):1563–1573. doi: 10.1038/mt.2009.123

Allele-specific RNAi Mitigates Phenotypic Progression in a Transgenic Model of Alzheimer's Disease

Edgardo Rodríguez-Lebrón 1, Cynthia M Gouvion 2, Steven A Moore 3, Beverly L Davidson 4, Henry L Paulson 1
PMCID: PMC2835271  PMID: 19532137

Abstract

Despite recent advances suggesting new therapeutic targets, Alzheimer's disease (AD) remains incurable. Aberrant production and accumulation of the Aβ peptide resulting from altered processing of the amyloid precursor protein (APP) is central to the pathogenesis of disease, particularly in dominantly inherited forms of AD. Thus, modulating the production of APP is a potential route to effective AD therapy. Here, we describe the successful use of an allele-specific RNA interference (RNAi) approach targeting the Swedish variant of APP (APPsw) in a transgenic mouse model of AD. Using recombinant adeno-associated virus (rAAV), we delivered an anti-APPsw short-hairpin RNA (shRNA) to the hippocampus of AD transgenic mice (APP/PS1). In short- and long-term transduction experiments, reduced levels of APPsw transprotein were observed throughout targeted regions of the hippocampus while levels of wild-type murine APP remained unaltered. Moreover, intracellular production of transfer RNA (tRNA)-valine promoter–driven shRNAs did not lead to detectable neuronal toxicity. Finally, long-term bilateral hippocampal expression of anti-APPsw shRNA mitigated abnormal behaviors in this mouse model of AD. The difference in phenotype progression was associated with reduced levels of soluble Aβ but not with a reduced number of amyloid plaques. Our results support the development of allele-specific RNAi strategies to treat familial AD and other dominantly inherited neurodegenerative diseases.

Introduction

Alzheimer's disease (AD) is the most common neurodegenerative disease and the leading cause of dementia in aging human populations. Although most AD occurs sporadically, studies of familial forms of AD have revealed that altered processing of amyloid precursor protein (APP) is the proximal molecular event leading to hallmarks of AD: amyloid deposition, progressive memory impairment, and cognitive dysfunction.1,2,3 A widely expressed membrane protein, APP is first cleaved by α- or β-secretase (BACE1), which generates the N-terminal, soluble α- or β-ectodomain (APPs-α/β), and membrane-bound C-terminal fragments. C-terminal fragments are then cleaved by γ-secretase to produce the APP intracellular domain and one of three major Aβ peptides of 40, 42, or 43 amino acids. Production and accumulation of Aβ plays a key role in AD pathogenesis through its detrimental effects on neuronal function and survival.4,5,6 Accordingly, several new therapeutic strategies have aimed to curtail the production or enhance the clearance of Aβ from the AD brain.7,8

RNA interference (RNAi) is a powerful tool to silence the expression of deleterious disease genes in the brain. Using viral and nonviral platforms, we and others have shown the feasibility of RNAi as a therapeutic approach to neurodegenerative diseases including AD.9 Initial RNAi studies in AD have focused on the genes encoding BACE1 and presenilin-1 (a component of the γ-secretase complex) as potential therapeutic targets because inhibiting their expression is predicted to reduce levels of Aβ. BACE1 is a desirable target because BACE1 null mice are viable and do not generate Aβ when crossed to transgenic mouse models of AD.10,11 Proof-of-principle RNAi experiments demonstrated that lentiviral delivery of short-hairpin RNAs (shRNAs) targeting BACE1 expression can improve neuropathology and behavioral deficits in a mouse model of AD.12 Recent studies, however, have shown that ablation of BACE1 leads to postnatal hippocampus-dependent deficits associated with altered NRG1 processing and signaling.13,14 Although it is not yet clear whether these phenotypes reflect developmental abnormalities or a need for BACE1 activity in the adult brain, they underscore the need for testing additional RNAi targets as potential therapy for AD.

APP is one such RNAi target. Variations in APP gene dosage directly correlate with Aβ production in the brain and constitute an important risk factor for AD. In Down syndrome, for example, overproduction of APP due to the extra copy of the APP gene at 21q21 leads to early-onset AD pathology.15 Genomic duplication of the APP locus also causes autosomal dominant early-onset AD with cerebral amyloid angiopathy, and allelic variants in the 5′ regulatory region of the APP gene that enhance APP transcription are associated with increased risk of AD.16,17 Importantly, age of onset in carriers of familial APP missense mutations is inversely correlated to levels of APP expression.18 Thus, mild but sustained reduction in the levels of brain APP would be predicted to have therapeutic benefit in AD.

We and others have taken advantage of the exquisite sequence specificity of RNAi to suppress the expression of toxic genes in an allele-specific manner.19 This strategy holds promise as an RNAi-based therapy for autosomal dominant neurological diseases where maintaining activity of the wild-type protein is important for normal central nervous system function. In the case of early-onset familial AD, autosomal dominant mutations in APP, PSEN1, and PSEN2 share as a common toxic mechanism the enhanced production of Aβ. Allele-specific RNAi targeting of these dominantly acting allelic variants thus may be beneficial in early-onset familial AD. We have previously demonstrated in cultured cells that siRNAs and shRNAs can be designed that specifically suppress either wild-type APP or an early-onset familial AD mutant APP allele, the Swedish (K670N/M671L) dinucleotide missense mutation (APPsw).20 Here, we designed experiments to determine whether viral delivery of these shRNAs mediates silencing of APP or APPsw in vivo and mitigates Aβ-induced pathology in a double-transgenic mouse model of AD. We found that recombinant adeno-associated virus pseudotype-5 (rAAV-2/5)–mediated delivery of shRNA targeting APPsw results in long-term, allele-specific suppression of APPsw expression in the hippocampus of AD mice. Importantly, transfer RNA (tRNA)-valine promoter–driven shRNAs did not suppress wild-type mouse APP or induce overt neuronal toxicity. Finally, progression of AD-like behavior and pathology was impaired by the injected shRNA. Our results support further development of RNAi-based gene transfer strategies for the treatment of AD and other neurodegenerative diseases.

Results

In vitro allele-specific silencing of APPsw

Efficient strategies have been devised that can achieve allele-specific silencing of dominant mutant alleles in vitro. Recent genetic studies using RNAi transgenic mice have further shown the feasibility of long-term allele-specific silencing in vivo.21 We previously identified an shRNA sequence that preferentially targeted the tandem dinucleotide missense APPsw mutation.20 Our strategy placed the dinucleotide mismatch between wild-type and mutant APP alleles at the central nucleotide positions of the predicted guide strand. In this strategy, the double purine–purine mismatch between wild-type APP mRNA and the guide strand targeting mutant APPsw straddles the predicted RNA-induced silencing complex cleavage site, impairing RNA-induced silencing complex cleavage of wild-type APP (Supplementary Figure S1).

To test the viability of a viral-mediated allele-specific RNAi strategy, we inserted tRNA-valine promoter–driven shRNAs against wild-type APP or APPsw into rAAV vectors. The activity of these rAAV vectors was initially assessed in vitro using COS-7 cells. We performed transient co-transfection of APP or APPsw expressing plasmids together with a negative control “shRNA-empty vector” expressing only green fluorescent protein (GFP) or with a vector encoding: (i) shRNA targeting a shared common sequence between APP and APPsw (shAPP), (ii) shRNA selectively targeting APPsw (shAPPsw), or (iii) shRNA directed against a GFP scrambled sequence as a negative control (shMiss). As expected, co-transfection of the shAPP plasmid led to a significant reduction in the levels of both wild-type APP and APPsw (~3% of control levels in both cases) when compared to the shMiss and GFP-only control vectors (Figure 1). In contrast, co-transfection of shAPPsw plasmid at a shRNA:target ratio of either 2:1 or 4:1 significantly suppressed APPsw expression (~76 and 12% of control levels, respectively), while wild-type APP levels remained largely unchanged at both target ratios (Figure 1). This preferential silencing in an allele-selective manner agrees with our previously published observations using similar shRNA targeting sequences.20

Figure 1.

Figure 1

Open in a new tab

Allele-specific silencing of APPsw in vitro. An shRNA vector targeting the dinucleotide mutation in APPsw (shAPPsw) significantly suppresses APPsw expression at two different shRNA to target ratios (b; lanes 5 and 6). In contrast, wild-type APP protein levels remain largely unaffected by coexpression of the same shAPPsw vector (a; lanes 5 and 6). An shRNA (shAPP) targeting an APP sequence outside the Swedish mutation effectively suppresses expression of both APP and APPsw (a, b; lanes 3 and 4). A GFP-only vector (a, b; lane 1) and a missense shRNA targeting a scrambled GFP sequence (a, b; lane 2) were used as transfection and nonspecific RNAi controls, respectively. Following detection of APP and APPsw with 6E10 antibody (top), blots were stripped and probed with α-tubulin antibody (bottom). (c) Quantification of three independent experiments. Digital images of APP, APPsw, and tubulin blots were generated and quantified using National Institutes of Health ImageJ software. Levels of APP and APPsw were normalized to those of tubulin (loading control) and expressed as percent of control (GFP-only vector). Numbers 2 to 1 and 4 to 1 in c refer to shRNA to target plasmid molar ratios. Error bars = mean ± SD. Unpaired t-test; *P < 0.0001. APP, amyloid precursor protein; APPsw, Swedish variant of APP; GFP, green fluorescent protein; RNAi, RNA interference; shRNA, short-hairpin RNA; Tub, tubulin.

Allele-specific silencing of APPsw in vivo

We next evaluated the efficacy of the tRNA-valine promoter–driven shRNA constructs in vivo, using a double-transgenic mouse model of AD that expresses both APPsw and a mutant form of presenilin-1 lacking exon 9 (hereafter referred to as AD transgenic).22,23 In this model, expression of the two transgenes is independently regulated by the mouse prion promoter. Importantly, in the protein-coding region targeted by shAPPsw, the only difference in nucleotide sequence between the mouse/human chimera APPsw transgene and endogenous murine APP is the tandem dinucleotide missense change in the Swedish mutation. Therefore, this mouse model of AD allowed us to study an allele-specific RNAi strategy in vivo.

We generated high-titer, pseudotyped rAAV-2/5 vectors encoding either shAPPsw (rAAV-shAPPsw) or shMiss (rAAV-shMiss). Two unilateral injections delivered a total of 2 µl (1 µl per site) of rAAV-shAPPsw or rAAV-shMiss into the hippocampus of 2- to 4-month-old wild-type (n = 5) or AD transgenic mice (n = 5). A humanized renilla GFP (hrGFP) reporter gene included in the rAAV vector (Figure 2a) allowed us to identify and evaluate the extent of rAAV transduction in the hippocampal formation. As shown in Figure 2b, 5 weeks postsurgery, robust hrGFP expression was observed in pyramidal neurons of the CA1 and CA2 regions in the injected hemisphere of wild-type or AD transgenic mice. We should emphasize that even though the hrGFP signal was evident in fibers extending to the contralateral uninjected hemisphere, very few hrGFP-positive cells were observed in the contralateral uninjected hippocampus. Thus, for subsequent analyses we employed contralateral injections of the control virus, rAAV-shMiss, to control for the effects of rAAV-shAPPsw.

Figure 2.

Figure 2

Open in a new tab

Efficient rAAV5-mediated transduction of hippocampal neurons. (a) tRNA-valine promoter–driven shRNAs (shAPPsw and shMiss) were cloned into a rAAV-2 inverted terminal repeat (iTR)–containing plasmid upstream of a CMV-driven, hrGFP expression cassette. Pseudotyped rAAV-2/5 virus was produced and injected into the hippocampus of wild-type or AD transgenic mice stereotactically. (b) Five weeks postsurgery, uniform and widespread hrGFP expression was observed throughout the hippocampus of young (2–4 months of age) AD mice. (c–e) By confocal immunofluorescence most NeuN-positive neurons (c) were transduced with AAV-RNAi vector as evidenced by colocalization of hrGFP (d), and NeuN signals (e). b, bar = 400 µm; c–e, bar = 150 µm. AD, Alzheimer's disease; CMV, cytomegalovirus; hrGFP, humanized renilla GFP; rAAV, recombinant adeno-associated virus; RNAi, RNA interference; shRNA, short-hairpin RNA; tVal, tRNA-valine.

rAAV vectors are commonly used as gene delivery vehicles in the central nervous system because they lack immunogenicity and are highly neurotropic.24 Consistent with this, hrGFP-positive regions of the hippocampus (CA1, CA2) stained with hematoxylin/eosin or NeuN antibody were morphologically indistinguishable from untransduced or uninjected controls (Supplementary Figure S2). High power confocal microscopy analysis of sections obtained from wild-type or AD transgenic mice stained with NeuN antibody confirmed both a lack of overt neuronal toxicity and a preference for neuronal transduction by both rAAV-shAPPsw and rAAV-shMiss (Figure 2c–e and Supplementary Figure S2).

To study the effects of rAAV-shAPPsw on transgene expression, we performed immunoblot and immunohistochemical analyses with the 6E10 monoclonal antibody, which recognizes the chimeric APPsw transprotein but not endogenous murine APP.25 In the hippocampus of injected AD mice, rAAV-shAPPsw significantly reduced levels of APPsw immunoreactivity throughout transduced regions, compared to contralateral hippocampi expressing control rAAV-shMiss (Supplementary Figure S3). Confocal imaging revealed that APPsw protein levels were markedly reduced in transduced (i.e., hrGFP-positive) pyramidal neurons expressing rAAV-shAPPsw. The diffuse and punctate APPsw staining in AD mice is nearly absent in neurons transduced with shAPPsw virus. In contrast, adjacent nontransduced neurons in the hippocampus retained full APPsw expression (Figure 3e–h). Moreover, there was no detectable loss of 6E10 immunoreactivity in neurons expressing rAAV-shMiss (Figure 3a–d).

Figure 3.

Figure 3

Open in a new tab

RNAi-mediated suppression of APPsw expression in vivo. (a–h) Confocal images of hippocampus from AD transgenic mice 5 weeks postinjection. Coronal sections containing hippocampus were obtained from injected mice, immunostained with anti-human APP antibody, 6E10 (red, a, e), and analyzed for hrGFP expression (green, b, f). Panels a–d are from an AD transgenic mouse injected with control rAAV-shMiss virus while panels e–h are from an AD transgenic mouse injected with rAAV-shAPPsw virus. (c, g) Merged images from a, b and e, f. (d, h) Enlargements of the boxed areas in c and g, respectively. In panel d, arrows label transduced (GFP-positive) hippocampal neurons that continue to express APPsw transprotein. By contrast, in panel h, GFP-positive hippocampal neurons show markedly reduced APPsw immunostaining (arrows) while adjacent nontransduced neurons retain robust APPsw expression (arrowheads). Images shown are representative of results from all analyzed mouse tissue (n = 5 mice per group). Bar = 50 µm. AD, Alzheimer's disease; APPsw, Swedish variant of APP; hrGFP, humanized renilla GFP; rAAV, recombinant adeno-associated virus.

To determine whether rAAV-shAPPsw expression resulted in allele-specific silencing of APPsw in vivo, a second set of unilateral injections was performed in young (4 months) AD transgenic mice (n = 10). In this experiment, mice received two unilateral injections (1 µl per site) of rAAV-shAPPsw into the right hippocampus and two injections (1 µl per site) of the control virus, rAAV-shMiss, into the contralateral left hippocampus. Five weeks after surgery, brains were harvested and fresh-frozen sections obtained. Transduced areas of the hippocampus (Figure 4a) were identified, quickly dissected, and total RNA or protein contents extracted. We designed “allele-specific” primers to compare the mRNA levels of endogenous APP and transgenic APPsw using real-time quantitative PCR. As shown in Figure 4b, following rAAV-shAPPsw expression the mRNA levels of APPsw were suppressed to <30% (unpaired t-test; P < 0.05) of those observed after expression of rAAV-shMiss in the hippocampus of AD transgenic mice. In contrast, the mRNA levels of endogenous murine APP were unaffected by the expression of rAAV-shAPPsw when compared to the rAAV-shMiss control. Western blot analysis confirmed allele-specific suppression at the protein level: there was a >50% loss (unpaired t-test; P < 0.05) of APPsw transprotein in areas of hippocampus transduced with rAAV-shAPPsw, when compared to areas transduced with control rAAV-shMiss virus (Figure 4c–d). In contrast, total levels of murine APP remained unchanged in rAAV-shAPPsw and rAAV-shMiss treated mice. Together these results demonstrate that tRNA-valine promoter–driven shRNA that targets APPsw can mediate highly effective, allele-specific silencing of APPsw in vivo.

Figure 4.

Figure 4

Open in a new tab

In vivo allele-specific silencing of APPsw. Five AD mice (4 months of age) were unilaterally injected in the hippocampus with rAAV-shAPPsw virus and contralaterally with the control rAAV-shMiss virus. (a) Fresh-frozen sections (20 sections per mouse, 5 mice per group) containing transduced hippocampus (arrow) were dissected using a razor blade to separate hippocampus from the rest of the brain tissue. Total RNA (b) or protein extracts (c, d) were analyzed using either real-time quantitative PCR or western blot. (b) An allele-specific qPCR assay was used to compare the effects of control rAAV-shMiss (black bars) or rAAV-shAPPsw (white bars) on the levels of murine APP RNA (left graph) to those of the APPsw transgene RNA (right graph). APP and APPsw mRNA levels are shown as a percent of APP and APPsw levels found in rAAV-shMiss transduced tissue after being normalized to levels of mouse c-myc (internal control). Levels of APPsw mRNA in rAAV-shAPPsw injected hippocampus were significantly reduced compared to contralateral control hippocampus. (c) In rAAV-shAPPsw injected hippocampus, APPsw protein levels (top) were reduced while levels of murine APP (middle) were unchanged by western blot analysis. APPsw was detected with 6E10 antibody, murine APP with 22C11 antibody. Tubulin levels were used as loading control (bottom). Results for 2 of the 5 mice analyzed per group are shown in c. The number under the group heading (shMiss or shAPPsw) identifies individual subjects within each group. (d) Results for all mice from both groups are summarized. Quantification was done as described in Figure 1c and reported as a percent of APP and APPsw protein levels in AD mice injected with rAAV-shMiss (percent control). Error bars for b and d = mean ± SD. Unpaired t-test; *P < 0.01. APP, amyloid precursor protein; APPsw, Swedish variant of APP; rAAV, recombinant adeno-associated virus; qPCR, quantitative real-time PCR; Tub, tubulin.

rAAV-shAPPsw mitigates the progression of AD-like behavioral impairments

In this AD transgenic mouse model, the accumulation of Aβ peptides in medial temporal brain regions results in age-related impairment in learning and memory tasks. Amyloid deposits can be detected as early as 6 months of age but are most evident past 8 months of age (Supplementary Figure S4). We sought to investigate the effects of rAAV-mediated APPsw RNAi on Aβ-induced neuropathology. Six- to eight-month-old male AD transgenic (n = 20) and nontransgenic littermate controls (n = 9) received bilateral intrahippocampal injections of either rAAV-shMiss (WT-shMiss, n = 4; AD-shMiss, n =10) or rAAV-shAPPsw (WT-shAPPsw, n = 5; AD-shAPPsw, n = 10).

Four months postsurgery, all four treatment groups were examined in a spatial memory task. Mice were tested in a Barnes circular maze twice a day for 9 consecutive days followed by a probe trial on day 10. Using this task, we tested the ability of mice to rely on the use of spatial cues to escape from a brightly lit, elevated open platform. The time taken to locate the hidden escape tunnel and the number of incorrect searches or errors were recorded as measures of spatial learning. For the purposes of analysis, trial data were binned into three different sets (set 1 = days 1, 2, and 3; set 2 = days 4, 5, and 6; set 3 = days 7, 8, and 9). As expected, repeated measures analysis of variance (ANOVA) revealed that escape latencies were significantly decreased across trials for all groups, independent of genotype or RNAi treatment [F(2, 84) = 14.9, P < 0.001] (Figure 5a). AD mice expressing the control rAAV-shMiss displayed longer escape latencies during the second set of trials (days 4, 5, and 6) when compared to the AD-shAPPsw group, and this difference approached significance [ANOVA; F(1, 18) = 4.05, P = 0.06]. We also analyzed the number of incorrect searches. During the second set of trials, AD transgenic mice exhibited more incorrect searches than did wild-type littermate controls [ANOVA, F(1, 27) = 25.8, P < 0.001], an effect which was independent of treatment (Figure 5b). During the last set of trials, however, the number of errors made by AD transgenic mice injected with rAAV-shMiss virus was significantly higher than in AD transgenic mice injected with the shAPPsw virus [ANOVA, F(1, 18) = 25.7, P < 0.001]. Notably, there was no significant difference between the number of errors committed by AD mice expressing anti-APPsw shRNA and wild-type littermates transduced with either of the two RNAi viruses.

Figure 5.

Figure 5

Open in a new tab

Improved performance in the Barnes maze spatial-learning task. Four months after bilateral delivery of rAAV into the hippocampus, four groups of mice (WT-shMiss n = 4, WT-shAPPsw n = 5, AD-shMiss n = 10, AD-shAPPsw n = 10) were tested for 9 consecutive days followed by a probe test on day 10. Data were binned into three sets: days 1–3, days 4–6, and days 7–9. (a) Analysis of mean escape latencies as a function of time. (b) Video tracking software was used to quantify the number of errors made during each trial. Bin analysis showed that nontransgenic littermate controls performed better (fewer errors) than AD transgenic mice during days 4–6 irrespective of treatment. During the last set of trials, AD transgenic mice expressing rAAV-shMiss performed significantly worse than all other groups, including AD transgenic expressing rAAV-shAPPsw. (c) The distance (number of holes) between the escape hole and the first explored hole at the beginning of each trial was recorded and analyzed as an indirect measure of search strategy. From the initial to the last set of trials, AD transgenic mice injected with control rAAV-shMiss virus continued to begin their exploration, on average, further from the escape hole than AD transgenic mice expressing shAPPsw or nontransgenic littermates. Error bars = mean ± SEM. *,#P < 0.001, ANOVA. AD, Alzheimer's disease; ANOVA, analysis of variance; rAAV, recombinant adeno-associated virus; WT, wild type.

To determine whether this improved performance in rAAV-shAPPsw injected AD mice correlated with differences in search strategy, we analyze the distance between the escape hole and the initial hole searched by the mice as an indirect measure of strategy (Figure 5c). Again, there was a significant effect of genotype during the second set of trials: AD transgenic mice appeared to randomly choose which hole to explore first whereas wild-type littermate controls appeared to use a more spatial cues–based strategy irrespective of treatment [ANOVA, F(1, 27) = 27.5, P < 0.001]. During the last set of trials, however, both wild-type mice (transduced with control or experimental virus) and AD transgenic mice transduced with rAAV-shAPPsw limited their search field to holes within the quadrant containing the escape hole (mean = 3.8 holes away). In contrast, AD transgenic mice injected with control rAAV-shMiss virus continued to explore outside the escape hole quadrant (mean = 12.5 holes away) suggesting persistent reliance on a random-based strategy (≥10 holes). Together, these data suggest that sustained expression of shAPPsw in the hippocampus of AD transgenic mice improves performance in a spatial memory task.

Next we addressed whether similar improvements occurred in an unrelated memory-dependent task. AD transgenic mice expressing control or anti-APPsw shRNAs were challenged using the novel object recognition protocol. After habituation to the training arena, mice were allowed to explore two identical objects (object 1 and object 2) for a total of 10 minutes. Following a long retention interval (24 hours), object 2 was replaced by a novel object (object 3) in the arena and mice were allowed to explore object 1 and object 3 for a total of 5 minutes. Percent time spent exploring object 1 (tO1) versus object 2 (tO2) and object 1 (tO1) versus object 3 (tO3) was measured and analyzed as a function of preference. The total percent time spent exploring (tO1 + tO2 and tO1 + tO3) was used to control for anxiety or motor impairments during each session. During initial training there were no significant differences between AD transgenic mice expressing rAAV-shMiss or rAAV-shAPPsw in object preference or percent time exploring the objects (Figure 6a,b). Thus, both groups of mice were equally able and motivated to explore. After a retention interval of 24 hours, AD transgenic mice injected with AAV-shMiss failed to show a novel object preference, spending similar amounts of time exploring both objects (tO1 = 48.7, tO3 = 51.3; paired t-test, P = 0.59). In contrast, AD transgenic mice transduced with anti-APPsw shRNA AAV explored the novel object significantly more (tO1 = 34.4, tO3 = 65.6; paired t-test, P < 0.001) (Figure 6c). Taken together these data suggest that selectively targeting APPsw via RNAi can mitigate the progression of Aβ-induced hippocampal impairment in this mouse model of AD.

Figure 6.

Figure 6

Open in a new tab

Performance in the object recognition task. AD-shMiss (n = 10) and AD-shAPPsw (n = 10) mice were challenged using a hippocampal-dependent object recognition task protocol. After habituation, two identical objects (object 1 and object 2) were introduced into the testing arena and exploratory behavior was recorded for 10 minutes. Both groups spent similar time exploring the objects (a) and did not show a preference for either object (b) during the acquisition phase of the test. After a retention interval of 24 hours, a novel object was introduced (object 3) and exploratory behavior was recorded for a total of 5 minutes. (c) AD transgenic mice injected with rAAV-shAPPsw showed increased exploration of novel object 3 (>60% of time) while AD mice injected with control rAAV-shMiss did not show any preference for exploring novel object 3. Error bars = mean ± SD. Unpaired t-test; *P < 0.001. AD, Alzheimer's disease; rAAV, recombinant adeno-associated virus.

Expression of rAAV-shAPPsw lowers hippocampal Aβ42 levels

Increased levels of soluble Aβ42 correlate with toxicity in the rodent brain.3,26,27,28 We sought to determine whether our AAV-based allele-specific RNAi strategy resulted in lowered levels of Aβ42 in the hippocampus of injected AD mice. hrGFP-positive, fresh-frozen hippocampal sections were obtained from AD mice bilaterally transduced with either AAV-shMiss or AAV-shAPPsw, quickly dissected and analyzed using a sandwich Aβ42 enzyme-linked immunosorbent assay. We observed a 62% reduction in the level of soluble Aβ42 in mice expressing rAAV-shAPPsw (22.4 ng/g) when compared to control AD mice that received bilateral injections of rAAV-shMiss virus (58.6 ng/g) (Figure 7a). These results concur with the ~50% reduction in holo-APPsw protein observed in unilaterally injected mice.

Figure 7.

Figure 7

Open in a new tab

Reduced hippocampal Aβ42 levels in AD mice transduced with rAAV-shAPPsw. (a) Fresh-frozen sections containing hrGFP-positive hippocampus from AD mice injected bilaterally with either rAAV-shAPPsw (n = 6) or control rAAV-shMiss virus (n = 6) were homogenized, and soluble Aβ42 levels were measured using a sandwich ELISA protocol. As shown in a, long-term bilateral expression of rAAV-shAPPsw in the hippocampus of AD transgenic mice led to significant reductions in the levels of soluble Aβ42 when compared to levels after expression of control rAAV-shMiss. Soluble Aβ42 levels are presented as Aβ42 ng/g of tissue. (b) Amyloid burden was visualized using a Congo red staining protocol. (c) For each mouse (n = 6 each for AD-shMiss and AD-shAPPsw) a total of 30 fields within transduced hippocampus (five sections per mouse) were analyzed for number of plaques and percent area covered by plaques. Untransduced cortical fields were used to normalize between mice. Digital images were analyzed using the newCAST software (Visiopharm Integrator System). Error bars = mean ± SD. AD, Alzheimer's disease, ELISA, enzyme-linked immunosorbent assay; hrGFP, humanized renilla GFP; rAAV, recombinant adeno-associated virus.

As previously shown, 8-month-old AD transgenic mice possess Congo Red–positive amyloid plaques that increase in size and number as the mice age (Supplementary Figure S4). We wished to determine whether our allele-specific RNAi approach had lowered the amyloid burden typical of this AD mouse model. Serial coronal sections (five per mouse) containing transduced hippocampus were stained with Congo red dye or the 6E10 antibody. For each mouse (n = 6 for each AD-shMiss and AD-shAPPsw), a total of 30 hippocampal fields were digitally imaged and scored for number and area covered by plaques using the newCAST analysis software (Visiopharm Integrator System, www.visiopharm.com). We failed to detect a significant difference in the hippocampal area covered by plaques or the number of plaques in control injected AD mice (rAAV-shMiss) versus AD mice expressing shAPPsw (Figure 7). In contrast and as previously observed (Figure 3h), transduced neurons were largely devoid of intracellular 6E10 positive staining.

Discussion

Validation of RNAi as a potential therapy to slow or halt neurodegenerative disease progression is a high priority. Among the important issues in RNAi therapy is the development of allele-specific approaches in which mutations or disease-linked polymorphisms are selectively targeted to suppress dominant acting disease genes without inhibiting the corresponding wild-type alleles.19 Here, in studies targeting mutant APPsw expression in AD transgenic mice, we tested the therapeutic efficacy of viral-mediated, allele-specific RNAi for familial AD. rAAV-mediated delivery of a shRNA targeting the APPsw mutation suppressed expression of mutant APP without affecting wild-type murine APP levels. Long-term (4 month) expression of APPsw shRNA in the hippocampus of AD mice was well tolerated and mitigated phenotypic features of disease. These results support the use of RNAi as a therapeutic strategy for AD and establish that allele-specific approaches for hereditary neurodegenerative diseases are feasible.

AD mice injected bilaterally with rAAV-shAPPsw performed better in two hippocampus-dependent tasks: the Barnes maze and an object recognition test. The results suggest that bilateral suppression of APPsw expression in pyramidal neurons of the hippocampus is sufficient to improve phenotypic behavior in this model of AD (Supplementary Figure S5). RNAi-mediated suppression of behavioral deficits in AD transgenic mice correlated with lowered levels of soluble Aβ42 rather than with a reduced number or volume of amyloid deposits. This study does not include a rigorous analysis of soluble Aβ42 species (i.e., monomeric versus oligomeric). Thus, we can not speculate about the soluble Aβ42 species that might be responsible for the behavioral deficits observed in this mouse model of AD. Recent evidence supports the view that increased levels of Aβ42 mediate early detrimental changes in synaptic activity in mouse models and AD patients.4,29 Moreover, in mice with a low amyloid burden, impaired spatial learning is more closely associated with increased Aβ42 load than with amyloid plaque number.26,27,28,30 We should emphasize that both the delivery of rAAV-shAPPsw and our subsequent analysis of its efficacy occurred before what are considered late stages of disease in this AD mouse model (~16 months of age).22,28,31 It remains to be seen whether suppression of APPsw via RNAi in older AD mice, when neuropathological deficits are clearly established, will be of benefit. In future studies, we plan to determine the effects that late stage disease delivery of rAAV-shAPPsw virus has on Aβ42 load, amyloid plaque burden and impaired learning phenotypes. This would provide us with an opportunity to begin dissecting the seemingly different contributions made by Aβ42 load and amyloid plaque burden during the AD disease process. Taken together with other reports27,32,33 our results suggest that reducing the levels of Aβ, in this case by employing RNAi directed at APP, should be of therapeutic benefit in early stages of AD.

Intraneuronal accumulations of Aβ42 have been previously associated with behavioral impairments in models of AD.34,35 In our study, confocal analysis of sections stained with the 6E10 antibody revealed a pattern of compartmentalized accumulations, or puncta, of APPsw in neurons throughout the brain. Whether these puncta represent intraneuronal accumulations of holo-APPsw protein or of Aβ fragments is currently unknown. It should be noted, however, that intracellular accumulations of either likely would perturb neuronal protein homeostasis and affect axonal or vesicle transport. These accumulations were markedly reduced in pyramidal neurons transduced with rAAV-shAPPsw virus. Several current therapeutic strategies for AD focus on disrupting the formation of extracellular plaques or enhancing their clearance from the brain, and thus are not designed to prevent intracellular accumulation of Aβ peptide or APP.7,8 Combining such extracellular approaches with RNAi-based therapies that reduce expression of the pathogenic protein might prove to be highly effective against Aβ-mediated pathological changes in AD.

Previous studies using an SOD1 transgenic mouse genetically crossed to RNAi transgenic mice showed effective allele-specific silencing in the central nervous system.21 This study addresses allele-specific RNAi using the more therapeutically relevant, gene transfer approach. Allele-specific RNAi-mediated silencing can be achieved by disrupting Watson–Crick base pairing along the Ago-2 cleavage site.19 The targeted APPsw mutation results in tandem purine–purine mismatches between the wild-type and mutant alleles. Central placement of these “bulky” mismatches in the shRNA used in our study (shAPPsw) led to allele-specific silencing of APPsw. Fortunately, in the case of APPsw the targeted sequence surrounding the mutation site is conducive to the design of shRNAs with favorable thermodynamic properties. This will not be the case, however, for all dominantly acting alleles carrying nucleotide variations. Although this dependence on the thermodynamic properties of designed shRNAs might appear to limit the application of allele-specific RNAi, recent studies suggest that this potential limitation can be overcome. For example, Schwarz et al.36 showed that purine:purine mismatches in the 16th position of the guide strand also mediate highly effective allele-specific RNAi. Thus, by shifting the mismatch away from the 10th nucleotide position the size of the targeted region can be expanded in the 3′ direction of the mRNA, increasing the probability of obtaining shRNAs with favorable thermodynamic properties for a given allele-specific target.

Several reports have established that expression of shRNAs can be toxic; in some cases this toxicity is associated with the high transcriptional activity of H1 and U6 promoters and inefficient processing of first generation hairpin designs.37,38,39 We employed a different pol III–based vector containing a tRNA-valine promoter expression cassette. This study reports the use of a tRNA-valine promoter–driven shRNA in mammalian brain. Histological analysis of brain tissue after 16 weeks of sustained shRNA expression did not reveal overt toxic effects. In principle, high level expression, inefficient processing, or accumulation of exogenous shRNA transcripts could interfere with the regulation of endogenous miRNAs, resulting in cellular toxicity. Interestingly, others have shown the tRNA-valine promoter to be inherently less active than the U6 and H1 promoters,40,41,42 and shRNAs driven from this promoter accumulate to a lesser extent than do U6- or H1-derived transcripts in the cell nucleus.42 We have yet to determine whether these two properties of the tRNA-valine expression cassette underlie the apparent lack of overt toxicity in our study. Careful analysis of miRNA profiles and in vivo processing of the shRNAs employed here should help identify any biologically significant differences between this tRNA-valine cassette and the more commonly used H1 or U6 platforms.

Autosomal dominant mutations in APP are relatively rare. They account for <20% of early-onset familial AD cases and the overwhelming majority of AD is not caused by pathogenic mutations.43 Allele-specific RNAi targeting of mutant APP would not apply to the common, sporadic form of AD. But since Aβ peptide generation is central to the pathogenesis of all AD, APP represents a plausible RNAi target for all AD. Toward this possibility, it will be important to test whether nonallele specific suppression of wild-type APP in fully mature neurons is detrimental to central nervous system function. In addition, localized versus more widespread parenchymal rAAV-RNAi-mediated suppression of APP should be investigated. APP and its proteolytic products likely participate in normal neuronal function. Suppressing APP expression by RNAi could prove therapeutically beneficial if it substantially reduced Aβ load without significantly disrupting the potentially critical function(s) of APP in the adult brain. We should point out, however, that recent studies provide evidence of the strong genetic heterogeneity associated with late-onset forms of AD, by far the most common form of AD.44 As scientists identify the genetic components responsible for modulating early- and late-onset forms of AD, allele-specific RNAi-based therapies may prove to be a valuable weapon in the therapeutic arsenal against AD.

Materials and Methods

tVal-RNA expression vectors and AAV. The tRNA-valine promoter–driven shRNA constructs (shAPP, shAPPsw, and shMiss) were generated by PCR as previously described20 and cloned into the pCR 2.1-TOPO vector (Invitrogen, Carlsbad, CA). The targeting and loop sequence for shAPPsw and shMiss were previously published.20 The shAPP construct was generated following a similar protocol using a reverse primer with the following sequence (5′–3′): AAAAAA CCCATTTCCAGAAAGCCAAACCAAGCTTCGTTTGGCTTTCTGGAAATGGGCTTCGAACCGGGGACCTTTCG. These transcriptional units were excised with EcoRI and subcloned into a rAAV shuttle plasmid upstream of a cytomegalovirus-driven hrGFP reporter gene.

rAAV serotype 2/5 (AAV-2 inverted terminal repeats; AAV-5 viral capsid) vectors were produced by the University of Iowa Vector Core facility using an insect cell (Sf9)/baculovirus-based system as previously described.45 A two-step purification process consisting of an iodixanol gradient (15–60% wt/vol) followed by an ion exchange (MustangQ Acrodisc membranes; Pall, East Hills, NY) was used to generate purified, high-titer rAAV virus. rAAV titers (viral genomes/ml) were determined by qPCR.

In vitro testing of shRNA efficacy. Complementary DNA constructs encoding APP and APPsw mutant proteins were kindly provided by R. Scott Turner (University of Michigan, Ann Arbor, MI). Transient co-transfections of target and shRNAs into COS-7 cells were performed using Lipofectamine or Lipofectamine Plus (Invitrogen) reagents following the manufacturer's guidelines as previously described. Protein lysates were obtained 48 hours post-transfection and analyzed by western blotting using previously reported methods.20 APP expression was detected using either monoclonal antibody 22C11 (1:1,000; Millipore, Billerica, MA) or human specific 6E10 (1:1,000; Millipore). Tubulin levels, serving as a loading control, were detected using a monoclonal antibody against α-tubulin (1:10,000; Sigma, St Louis, MO). Following incubation with peroxidase-conjugated anti-mouse secondary antibodies (1:20,000; Jackson Immuno Research Laboratories, West Grove, PA) membranes were treated with an ECL-plus reagent (Western Lighting; PerkinElmer, Waltham, MA) and exposed to film.

Mouse husbandry. All animal protocols were approved by the Institutional Animal Care and Use Committee at the University of Iowa. APPsw/ PS1Δexon9 double-transgenic mice [B6C3-Tg (APPswe,PSEN1dE9) 85Dbo/J] and wild-type littermates were bred and maintained in the animal husbandry at the University of Iowa. All mice were given chow and water ad libitum and housed in groups of four in a controlled temperature environment following a 12-hour light and dark cycle.

Stereotaxic injections. Stereotaxic administration of rAAV vectors was performed on 4- to 8-month-old wild-type or double-transgenic APPsw/PS1Δexon9 mice placed under anesthesia using a mixture of ketamine/xylazine. Mice received intrahippocampal injections (two sites/hemisphere) of rAAV2/5 encoding for either shAPPsw or shMiss diluted in Lactated Ringer's solution. For each injection, a total of 1 µl/site was delivered at an infusion rate of 0.25 µl/minutes using a 10-µl Hamilton syringe retrofitted with a glass micropipette with an outer diameter of 60–80 µm. One minute after the infusion was completed the micropipette was retracted 0.5 mm and allowed to remain in place for an additional 4 minutes before its complete removal from the mouse brain. Anterior–posterior and medial–lateral coordinates were calculated from bregma and the dorsal–ventral coordinates were calculated from the dural surface. Coordinates used during this study for two sites in the same hemisphere were anterior–posterior: −1.9, medial–lateral: ±1.4, dorsal–ventral: −1.5 for site 1 and anterior–posterior: −2.1, medial–lateral: ±2.0, dorsal–ventral: −1.4 for site 2. These measurements were made on an experimentally determined flat skull.

In vivo testing of shRNA efficacy. Four-month-old APPsw/PS1d9 mice or nontransgenic littermates received unilateral injections (two sites) to determine the effect that rAAV-shAPPsw and rAAV-shMiss had on the levels of APPsw transprotein. In this series of experiments, the contralateral hemisphere received either a sham or rAAV-shMiss injection as an internal control. For immunohistochemical analyses mice were deeply anesthetized with a ketamine/xylazine mix and transcardially perfused with 0.9% saline, followed by an ice-cold 4% paraformaldehyde/0.1 mol/l PO4 buffer solution. Brains were dissected, postfixed at 4 °C overnight, cryoprotected in 30% sucrose, and serial 40-µm thick coronal sections were collected in a sliding microtome. For molecular studies, anesthetized mice were transcardially perfused with ice-cold 0.9% saline containing a cocktail of protease inhibitors (Sigma) and the brains were quickly removed and flash frozen in liquid nitrogen. Tissue sections (15 µm) were obtained and thaw mounted onto SuperFrost slides (Fisher Scientific, Pittsburgh, PA) before storage at −80 °C.

Free floating sections were processed using standard immunohistochemical protocols as previously described.46 Briefly, tissue sections containing hrGFP-positive anterior hippocampus were rinsed 3 × 10 minutes with phosphate-buffered saline (PBS) (100 mmol/l phosphate, pH 7.4, and 0.9% NaCl), incubated for 10 minutes in 0.1% (vol/vol) H2O2/10% methanol and rinsed 3 × 10 minutes in PBS. For the detection of APPsw transprotein, an antigen retrieval step (incubation at 80 °C for 20 minutes in citrate buffer, pH 6.0) was performed before blocking of nonspecific antigenic sites [0.1% (vol/vol) Triton X-100/PBS containing 5% normal goat serum]. Sections were incubated overnight at 4 °C in blocking buffer containing a 1:500 dilution of the 6E10 monoclonal antibody, rinsed 3 × 10 minutes in 0.1% Triton X-100/PBS and incubated in secondary antibody (1:250, Alexa Fluor 568 goat anti-mouse; Invitrogen) for 2 hours at room temperature. Finally, sections were rinsed 3 × 10 minutes in PBS before mounting onto SuperFrost slides. In some instances, sections were incubated in biotinylated goat anti-mouse secondary antibody (1:500) followed by incubation in ABC reagent (Vector Laboratories, Burlingame, CA) and detection using 3-3′-diaminobenzidine as a color substrate. Detection of NeuN was performed in the same manner as described above using a monoclonal mouse anti-NeuN antibody at a 1:500 dilution. Hematoxylin/eosin staining was done for gross morphological analysis of sections. Digital images were obtained using either an Olympus IX71 inverted microscope or an Olympus BX51 upright (Olympus, Tokyo, Japan).

Relative mRNA levels of both murine and transgenic APPsw were determined using quantitative real-time PCR. Briefly, RNA was obtained by Trizol (Invitrogen) extraction of hrGFP-positive hippocampi (20 frozen sections per animal). Following spectrophotometric quantification, 1 µg of total RNA was reverse-transcribed (iScript cDNA Synthesis Kit; BioRad, Hercules, CA) and the resulting cDNA subjected to quantitative real-time PCR analysis with gene specific primers for murine APP (fwd: ctccaagatgcagcagaacggat; rev: cgatgggtagtgaagcaatggt), APPsw transgene (fwd: ccaagatgcagcagaacggcta; rev: ccgcaagaatgagaaccacctc), and c-myc (fwd: tcaagaggtgccacgtctcc; rev: tcttggcagctggatagtcctt) using the iQ SYBR Green Supermix in a myiQ Single Color RTPCR system (BioRad). The levels of murine APP and APPsw transgene mRNA were normalized to those of c-myc for each sample run and expressed as percent of levels found in rAAV-shMiss transduced samples (% control).

For analysis of murine APP and APPsw transprotein levels, hrGFP-positive hippocampi were extracted with 100 µl of brain lysis buffer [50 mmol/l Tris, 150 mmol/l NaCl, 0.1% sodium dodecyl sulfate (w/v), 0.5% Na-deoxycholate (wt/vol), 1% NP-40 (vol/vol), pH 7.4]. Total protein (50 µg) was resolved on a 6% sodium dodecyl sulfate–polyacrylamide gel electrophoresis and transferred to a polyvinylidene fluoride membrane for western blot analysis. Membranes were blocked in 5% milk/0.05% Tween-20/Tris-buffered saline, blotted with either 22C11, 6E10 or α-tubulin antibodies overnight at 4 °C, incubated with a peroxidase-conjugated secondary antibody (1:10,000) for 1 hour at room temperature followed by development using the ECL-plus reagent before exposure to film as described above. Band intensities were calculated using ImageJ software (NIH, Bethesda, MD) and reported as a percent of levels found in control samples (rAAV-shMiss).

Behavioral assessments. To study the phenotypic effects of reduced APPsw transprotein expression, we performed bilateral injections (four sites total) of rAAV-shAPPsw or rAAV-shMiss into the hippocampus of 6.5- to 8-month-old APPsw/PS1d9 bigenic mice (n = 20) and wild-type littermate controls (n = 9). Four months postsurgery, all four groups were challenged with two memory-dependent behavioral tasks: Barnes maze and object recognition.

Barnes maze: We first measured spatial-learning using the Barnes maze escape paradigm. In this task, mice rely on the use of spatial cues to escape from a brightly lit elevated open platform. This task is considered to be less anxiogenic than water-based mazes, as it does not require an aversive stimulus.47,48 Others have demonstrated age-dependent learning impairments in mouse models of AD using this task. Our Barnes maze consisted of a white acrylic circular platform (2,642 cm2) with 40 evenly spaced holes located 51 cm from the center of the platform and 5.5 cm from the edge. The escape hole led to a dark acrylic box (44 cm × 11 cm × 8 cm) fitted to the underside of the white platform. Mice have to learn the spatial relationship between the escape hole and the visual cues placed around the testing room in order to successfully complete the task. The maze was balanced and elevated from the floor (138 cm) using a pivoting metal rod that allowed for 360° rotation. The “starter box” was a top and bottomless acrylic cylinder (1,358 cm3) wrapped in aluminum foil. All trials were recorded using a digital video camera connected to a PC and analyzed using the VideoTrack system from ViewPoint (Montreal, Canada). Software analyses included latency to escape, number of errors, path length to the escape hole, and locomotion (speed and total distance of travel).

Four months after surgery AD transgenic and wild-type littermates transduced with rAAV-shAPPsw or rAAV-shMiss received 9 days of consecutive training consisting of two trials per day. A probe trial was performed on day 10. For each trial, the mouse was placed inside the “starter box” and allowed to habituate for 1 minute. Video capture began after the mouse was permitted to explore the platform by lifting the “starter box”. Each mouse was allotted a maximum of 5 minutes to enter the escape hole. Mice that were successful in escaping within the allotted time were kept inside the escape box for 1 minute before returning them to their home cages. If the mouse did not enter the escape hole within the 5 minutes time limit, it was gently brought to the escape hole and encouraged to enter into the dark escape box where it remained for 1 minute before returning to its home cage. Trials for each mouse were completed before starting the first subject on its second trial. To prevent odors from being used as cues, all surfaces were thoroughly cleaned following each trial using a 70% alcohol solution.

Object recognition: Mice display a natural preference to explore a novel object more often than a familiar one. Others have reported that this memory-dependent behavior becomes progressively impaired in mouse models of AD.49,50 Seven days after the Barnes maze test, AD transgenic and wild-type littermates transduced with rAAV-shAPPsw or rAAV-shMiss were tested using the novel object recognition protocol. Mice were first habituated to the training arena (which included nonexperimental objects) for 3 days. During the testing period, each mouse was placed inside the training arena (25 cm × 40.5 cm × 50 cm) with two identical objects (object 1 and object 2) for a total of 10 minutes. Twenty-four hours post-training object 2 was replaced by a novel object (object 3) and mice were allowed to explore object 1 and object 3 during a 5-minute trial. The arena was thoroughly cleaned between trials using a 70% alcohol solution. Each session was recorded using a digital video camera connected to a PC and analyzed using the VideoTrack system from ViewPoint. Percent time spent exploring object 1 (tO1) versus object 2 (tO2) and object 1 (tO1) versus object 3 (tO3) was measured and analyzed as a function of preference. The total percent of time spent exploring (tO1 + tO2 and tO1 + tO3) was used to control for anxiety or motor impairments during each session.

Soluble Aβ42 enzyme-linked immunosorbent assay. Following behavior analyses, frozen sections from bilaterally injected mice (AD rAAV-shMiss n = 6, AD rAAV-shAPPsw n = 6) were used to analyze soluble Aβ42 levels in transduced regions of the hippocampus following standard protocols.28 Briefly, frozen sections containing hrGFP-positive hippocampus were dissected using a razor blade, the collected tissue was weighed and soluble protein was obtained by homogenization of samples in PBS, pH 7.2, containing a cocktail of protease inhibitors (Sigma). Two centrifugation steps were performed in order to clear out most insoluble debris. Levels of Aβ42 were measured using a BetaMark x-42 enzyme-linked immunosorbent assay (Covance, Princeton, NJ) following the manufacturer's recommendations. Soluble Aβ42 levels were reported as ng/g of tissue.

Analysis of plaque burden. Two different methods were used to score plaque deposition in APPsw/PS1 transgenic mice: Congo Red staining and the 6E10 monoclonal antibody. Immunohistochemical staining with 6E10 monoclonal antibody was described above. To calculate plaque burden a total of 30 fields (three fields on the right hippocampus, three fields on the left hippocampus; five sections per mouse) within transduced hippocampus were analyzed using the newCAST software for number of plaques and percent area covered by plaques. Analysis of plaque burden throughout untransduced cortical regions was used to control for the natural variability observed between mice. To determine the area covered by plaques, the approximate size (µm2) of each Congo red–positive plaque in the randomly chosen fields was determined and the total was divided by the area (µm2) of the analyzed region (percent area covered). Cortical areas were analyzed in a similar fashion. Hippocampus percent area coverage was then divided by the cortical percent area coverage and this ratio was reported as a percent of the AD rAAV-shMiss control. Number of plaques was calculated by counting Congo red (or 6E10)–positive plaques present within the analyzed fields (30 fields per mouse within transduced hippocampus), normalized to cortical counts, and reported as a percent of the counts found in the rAAV-shMiss control treated mice.

Statistical analysis. Statistical analyses were performed using the Statview (SAS Institute, San Francisco, CA) analytical program. Unpaired Student's t-test was used for the analysis of western blots, quantitative real-time PCR and to assess differences in the object recognition task. A P value <0.05 was used to reject the null hypothesis. A repeated measures ANOVA (Statview) was used to determine the probability of differences between experimental groups in the Barnes maze task (escape latency, number of errors, number of holes away from escape). Significant main effects at P < 0.05 were also tested by one-way ANOVA for independent samples and a Fisher's protected least significant difference post hoc analysis reported.

SUPPLEMENTARY MATERIALFigure S1. Design of allele-specific RNAi. Our RNAi targeting strategy placed the two nucleotide mismatch between wild-type APP and mutant APPsw alleles (GA-->UC) at the 10th and 11th nucleotide positions of the predicted shRNA guide strand (light gray font). Uninterrupted complimentary base pairing between APPsw mRNA (left column) and the APPsw shRNA guide strand results in RISC-dependent cleavage and degradation of target RNA. The double purine-purine mismatch between wild-type APP mRNA (right column) and the guide strand targeting mutant APPsw is predicted to disrupt RISC-dependent activity on wild-type APP mRNA. This disruption, likely due to the chemical “bulkiness” inherent to purine-purine interactions, spares the wild-type APP allele.Figure S2. Gross morphological analysis of AAV-mediated RNAi transduction. Coronal sections obtained from wild-type and AD transgenic mice injected with either rAAV-shAPPsw or rAAV-shMiss were analyzed for gross morphological abnormalities following short-term rAAV-RNAi expression. (a) Hematoxylin and eosin (H&E) staining of mouse hippocampus. Age-matched comparisons between injected (AAV shMiss and AAV shAPPsw) and uninjected control AD transgenic mice did not reveal gross differences in the number of neurons or hippocampal structure. (b) Analysis of NeuN expression throughout transduced hippocampus confirmed results obtained in a. Scale bar= 450μm.Figure S3. Efficient widespread silencing of APPsw expression. (a–h) Five weeks post-rAAV delivery, coronal sections were analyzed for both hrGFP and APPsw expression. rAAV-RNAi-mediated (shAPPsw and shMiss) transduction was most evident throughout the CA1 and CA2 regions of the anterior hippocampus as shown by hrGFP expression (a–d). Loss of APPsw expression (6E10 antibody) was most pronounced in areas of the hippocampus transduced with rAAV-shAPPsw virus (e, f). In contrast, contralateral expression of rAAV-shMiss had no effect on the expression pattern or intensity of APPsw transprotein in the contralateral hippocampus (g, h). Images presented are from two different mice and are representative of all other mice analyzed (n = 5). (i–n) To confirm this result, sections were subjected to immunohistochemical analysis using 3-3'diaminobenzidine as a color substrate. Again, we observed reduced levels of APPsw immunoreactivity throughout transduced regions of the hippocampus (i, j), compared to contralateral hippocampus expressing negative control virus, rAAV-shMiss (l, m). (k) Digital zoom of j showing the transduction boundary with a clearly demarcated difference in APPsw expression. (n) Digital zoom of m showing the presence and continuity of APPsw expression throughout the rAAV-shMiss transduced hippocampus. Scale bar = 250μm.Figure S4. Age-dependent increase in amyloid burden in AD transgenic mice (APPsw/PS1Δ9). Congo red dye was used to visualize plaque burden in the brains of AD transgenic mice that were 4, 6, 8 or 12 months of age. Congo red-positive plaques begin to appear in the hippocampus of AD mice by 6 months of age (middle panels) and continue to increase in size and number at 8 months of age (right panels). Four month-old transgenic mice did not have Congo red-positive deposits whereas 12 month-old transgenic presented with the most amyloid load of all ages tested (not shown). Wild-type littermate controls were devoid of Congo red staining. WT = wild-type littermates, Tg = AD transgenic, BF= bright field, CR= Congo red signal. Scale bar = 250μm.Figure S5. Suppression of APPsw levels in bilaterally rAAV-shAPPsw injected AD transgenic mice (APPsw/PS1Δ9). Total protein was extracted from transduced hippocampi and analyzed for transgene expression using the 6E10 antibody as described in figure 4. Long term expression of rAAV-shAPPsw in the hippocampus of AD transgenic mice (lanes 2 and 3) led to a significant reduction in the levels of APPsw protein when compared to rAAV-shMiss control injected (lanes 5 and 6) or uninjected age-matched AD mice (lanes 1 and 4). Tubulin levels were used as loading control (bottom). The number under the group heading (rAAV-shAPPsw or rAAV-shMiss) identifies individual subjects within each group.

Supplementary Material

Figure S1.

Design of allele-specific RNAi. Our RNAi targeting strategy placed the two nucleotide mismatch between wild-type APP and mutant APPsw alleles (GA-->UC) at the 10th and 11th nucleotide positions of the predicted shRNA guide strand (light gray font). Uninterrupted complimentary base pairing between APPsw mRNA (left column) and the APPsw shRNA guide strand results in RISC-dependent cleavage and degradation of target RNA. The double purine-purine mismatch between wild-type APP mRNA (right column) and the guide strand targeting mutant APPsw is predicted to disrupt RISC-dependent activity on wild-type APP mRNA. This disruption, likely due to the chemical “bulkiness” inherent to purine-purine interactions, spares the wild-type APP allele.

Click here for supplemental data (8.7MB, tiff)
Figure S2.

Gross morphological analysis of AAV-mediated RNAi transduction. Coronal sections obtained from wild-type and AD transgenic mice injected with either rAAV-shAPPsw or rAAV-shMiss were analyzed for gross morphological abnormalities following short-term rAAV-RNAi expression. (a) Hematoxylin and eosin (H&E) staining of mouse hippocampus. Age-matched comparisons between injected (AAV shMiss and AAV shAPPsw) and uninjected control AD transgenic mice did not reveal gross differences in the number of neurons or hippocampal structure. (b) Analysis of NeuN expression throughout transduced hippocampus confirmed results obtained in a. Scale bar= 450μm.

Click here for supplemental data (42.9MB, tiff)
Figure S3.

Efficient widespread silencing of APPsw expression. (a–h) Five weeks post-rAAV delivery, coronal sections were analyzed for both hrGFP and APPsw expression. rAAV-RNAi-mediated (shAPPsw and shMiss) transduction was most evident throughout the CA1 and CA2 regions of the anterior hippocampus as shown by hrGFP expression (a–d). Loss of APPsw expression (6E10 antibody) was most pronounced in areas of the hippocampus transduced with rAAV-shAPPsw virus (e, f). In contrast, contralateral expression of rAAV-shMiss had no effect on the expression pattern or intensity of APPsw transprotein in the contralateral hippocampus (g, h). Images presented are from two different mice and are representative of all other mice analyzed (n = 5). (i–n) To confirm this result, sections were subjected to immunohistochemical analysis using 3-3'diaminobenzidine as a color substrate. Again, we observed reduced levels of APPsw immunoreactivity throughout transduced regions of the hippocampus (i, j), compared to contralateral hippocampus expressing negative control virus, rAAV-shMiss (l, m). (k) Digital zoom of j showing the transduction boundary with a clearly demarcated difference in APPsw expression. (n) Digital zoom of m showing the presence and continuity of APPsw expression throughout the rAAV-shMiss transduced hippocampus. Scale bar = 250μm.

Click here for supplemental data (44.9MB, tiff)
Figure S4.

Age-dependent increase in amyloid burden in AD transgenic mice (APPsw/PS1Δ9). Congo red dye was used to visualize plaque burden in the brains of AD transgenic mice that were 4, 6, 8 or 12 months of age. Congo red-positive plaques begin to appear in the hippocampus of AD mice by 6 months of age (middle panels) and continue to increase in size and number at 8 months of age (right panels). Four month-old transgenic mice did not have Congo red-positive deposits whereas 12 month-old transgenic presented with the most amyloid load of all ages tested (not shown). Wild-type littermate controls were devoid of Congo red staining. WT = wild-type littermates, Tg = AD transgenic, BF= bright field, CR= Congo red signal. Scale bar = 250μm.

Click here for supplemental data (10.1MB, tiff)
Figure S5.

Suppression of APPsw levels in bilaterally rAAV-shAPPsw injected AD transgenic mice (APPsw/PS1Δ9). Total protein was extracted from transduced hippocampi and analyzed for transgene expression using the 6E10 antibody as described in figure 4. Long term expression of rAAV-shAPPsw in the hippocampus of AD transgenic mice (lanes 2 and 3) led to a significant reduction in the levels of APPsw protein when compared to rAAV-shMiss control injected (lanes 5 and 6) or uninjected age-matched AD mice (lanes 1 and 4). Tubulin levels were used as loading control (bottom). The number under the group heading (rAAV-shAPPsw or rAAV-shMiss) identifies individual subjects within each group.

Click here for supplemental data (9.5MB, tiff)

Acknowledgments

We thank Adam Mallenger and Tina Knutson for their expert technical assistance, Pedro Gonzalez-Alegre and Sokol Todi for their critical comments on the manuscript, and members of the Paulson lab for their continued support. This project was funded by the National Institutes of Health (NS-50210 and a post-doctoral minority supplement to NS-50210) and the University of Michigan Alzheimer Disease Research Center (Pilot Grant P50-AG008761). This work was done at Iowa City, Iowa, United States and Ann Arbor, Michigan, United States.

REFERENCES

  1. Bertram L., and , Tanzi RE. Thirty years of Alzheimer's disease genetics: the implications of systematic meta-analyses. Nat Rev Neurosci. 2008;9:768–778. doi: 10.1038/nrn2494. [DOI] [PubMed] [Google Scholar]
  2. Thinakaran G., and , Koo EH. Amyloid precursor protein trafficking, processing, and function. J Biol Chem. 2008;283:29615–29619. doi: 10.1074/jbc.R800019200. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Selkoe DJ. Soluble oligomers of the amyloid β-protein impair synaptic plasticity and behavior. Behav Brain Res. 2008;192:106–113. doi: 10.1016/j.bbr.2008.02.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Shankar GM, Li S, Mehta TH, Garcia-Munoz A, Shepardson NE, Smith I, et al. Amyloid-β protein dimers isolated directly from Alzheimer's brains impair synaptic plasticity and memory. Nat Med. 2008;14:837–842. doi: 10.1038/nm1782. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Small SA., and , Duff K. Linking Aβ and tau in late-onset Alzheimer's disease: a dual pathway hypothesis. Neuron. 2008;60:534–542. doi: 10.1016/j.neuron.2008.11.007. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Lesné S, Koh MT, Kotilinek L, Kayed R, Glabe CG, Yang A, et al. A specific amyloid-β protein assembly in the brain impairs memory. Nature. 2006;440:352–357. doi: 10.1038/nature04533. [DOI] [PubMed] [Google Scholar]
  7. Brody DL., and , Holtzman DM. Active and passive immunotherapy for neurodegenerative disorders. Annu Rev Neurosci. 2008;31:175–193. doi: 10.1146/annurev.neuro.31.060407.125529. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Spencer B, Rockenstein E, Crews L, Marr R., and , Masliah E. Novel strategies for Alzheimer's disease treatment. Expert Opin Biol Ther. 2007;7:1853–1867. doi: 10.1517/14712598.7.12.1853. [DOI] [PubMed] [Google Scholar]
  9. Boudreau RL., and , Davidson BL. RNAi therapy for neurodegenerative diseases. Curr Top Dev Biol. 2006;75:73–92. doi: 10.1016/S0070-2153(06)75003-7. [DOI] [PubMed] [Google Scholar]
  10. Ohno M, Sametsky EA, Younkin LH, Oakley H, Younkin SG, Citron M, et al. BACE1 deficiency rescues memory deficits and cholinergic dysfunction in a mouse model of Alzheimer's disease. Neuron. 2004;41:27–33. doi: 10.1016/s0896-6273(03)00810-9. [DOI] [PubMed] [Google Scholar]
  11. Roberds SL, Anderson J, Basi G, Bienkowski MJ, Branstetter DG, Chen KS, et al. BACE knockout mice are healthy despite lacking the primary β-secretase activity in brain: implications for Alzheimer's disease therapeutics. Hum Mol Genet. 2001;10:1317–1324. doi: 10.1093/hmg/10.12.1317. [DOI] [PubMed] [Google Scholar]
  12. Singer O, Marr RA, Rockenstein E, Crews L, Coufal NG, Gage FH, et al. Targeting BACE1 with siRNAs ameliorates Alzheimer disease neuropathology in a transgenic model. Nat Neurosci. 2005;8:1343–1349. doi: 10.1038/nn1531. [DOI] [PubMed] [Google Scholar]
  13. Wang H, Song L, Laird F, Wong PC., and , Lee HK. BACE1 knock-outs display deficits in activity-dependent potentiation of synaptic transmission at mossy fiber to CA3 synapses in the hippocampus. J Neurosci. 2008;28:8677–8681. doi: 10.1523/JNEUROSCI.2440-08.2008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Savonenko AV, Melnikova T, Laird FM, Stewart KA, Price DL., and , Wong PC. Alteration of BACE1-dependent NRG1/ErbB4 signaling and schizophrenia-like phenotypes in BACE1-null mice. Proc Natl Acad Sci USA. 2008;105:5585–5590. doi: 10.1073/pnas.0710373105. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Isacson O, Seo H, Lin L, Albeck D., and , Granholm AC. Alzheimer's disease and Down's syndrome: roles of APP, trophic factors and ACh. Trends Neurosci. 2002;25:79–84. doi: 10.1016/s0166-2236(02)02037-4. [DOI] [PubMed] [Google Scholar]
  16. Brouwers N, Sleegers K, Engelborghs S, Bogaerts V, Serneels S, Kamali K, et al. Genetic risk and transcriptional variability of amyloid precursor protein in Alzheimer's disease Brain 20061292984–2991.Pt 11 [DOI] [PubMed] [Google Scholar]
  17. Sleegers K, Brouwers N, Gijselinck I, Theuns J, Goossens D, Wauters J, et al. APP duplication is sufficient to cause early onset Alzheimer's dementia with cerebral amyloid angiopathy Brain 20061292977–2983.Pt 11 [DOI] [PubMed] [Google Scholar]
  18. Theuns J, Brouwers N, Engelborghs S, Sleegers K, Bogaerts V, Corsmit E, et al. Promoter mutations that increase amyloid precursor-protein expression are associated with Alzheimer disease. Am J Hum Genet. 2006;78:936–946. doi: 10.1086/504044. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Rodriguez-Lebron E., and , Paulson HL. Allele-specific RNA interference for neurological disease. Gene Ther. 2006;13:576–581. doi: 10.1038/sj.gt.3302702. [DOI] [PubMed] [Google Scholar]
  20. Miller VM, Gouvion CM, Davidson BL., and , Paulson HL. Targeting Alzheimer's disease genes with RNA interference: an efficient strategy for silencing mutant alleles. Nucleic Acids Res. 2004;32:661–668. doi: 10.1093/nar/gkh208. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Xia X, Zhou H, Huang Y., and , Xu Z. Allele-specific RNAi selectively silences mutant SOD1 and achieves significant therapeutic benefit in vivo. Neurobiol Dis. 2006;23:578–586. doi: 10.1016/j.nbd.2006.04.019. [DOI] [PubMed] [Google Scholar]
  22. Jankowsky JL, Fadale DJ, Anderson J, Xu GM, Gonzales V, Jenkins NA, et al. Mutant presenilins specifically elevate the levels of the 42 residue β-amyloid peptide in vivo: evidence for augmentation of a 42-specific γ-secretase. Hum Mol Genet. 2004;13:159–170. doi: 10.1093/hmg/ddh019. [DOI] [PubMed] [Google Scholar]
  23. Jankowsky JL, Slunt HH, Ratovitski T, Jenkins NA, Copeland NG., and , Borchelt DR. Co-expression of multiple transgenes in mouse CNS: a comparison of strategies. Biomol Eng. 2001;17:157–165. doi: 10.1016/s1389-0344(01)00067-3. [DOI] [PubMed] [Google Scholar]
  24. Mandel RJ, Manfredsson FP, Foust KD, Rising A, Reimsnider S, Nash K, et al. Recombinant adeno-associated viral vectors as therapeutic agents to treat neurological disorders. Mol Ther. 2006;13:463–483. doi: 10.1016/j.ymthe.2005.11.009. [DOI] [PubMed] [Google Scholar]
  25. Thakker DR, Weatherspoon MR, Harrison J, Keene TE, Lane DS, Kaemmerer WF, et al. Intracerebroventricular amyloid-β antibodies reduce cerebral amyloid angiopathy and associated micro-hemorrhages in aged Tg2576 mice. Proc Natl Acad Sci USA. 2009;106:4501–4506. doi: 10.1073/pnas.0813404106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Cheng IH, Scearce-Levie K, Legleiter J, Palop JJ, Gerstein H, Bien-Ly N, et al. Accelerating amyloid-β fibrillization reduces oligomer levels and functional deficits in Alzheimer disease mouse models. J Biol Chem. 2007;282:23818–23828. doi: 10.1074/jbc.M701078200. [DOI] [PubMed] [Google Scholar]
  27. Lesné S, Kotilinek L., and , Ashe KH. Plaque-bearing mice with reduced levels of oligomeric amyloid-β assemblies have intact memory function. Neuroscience. 2008;151:745–749. doi: 10.1016/j.neuroscience.2007.10.054. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Puoliväli J, Wang J, Heikkinen T, Heikkilä M, Tapiola T, van Groen T, et al. Hippocampal Aβ42 levels correlate with spatial memory deficit in APP and PS1 double transgenic mice. Neurobiol Dis. 2002;9:339–347. doi: 10.1006/nbdi.2002.0481. [DOI] [PubMed] [Google Scholar]
  29. Walsh DM., and , Selkoe DJ. Aβ oligomers – a decade of discovery. J Neurochem. 2007;101:1172–1184. doi: 10.1111/j.1471-4159.2006.04426.x. [DOI] [PubMed] [Google Scholar]
  30. Reiserer RS, Harrison FE, Syverud DC., and , McDonald MP. Impaired spatial learning in the APPSwe + PSEN1DeltaE9 bigenic mouse model of Alzheimer's disease. Genes Brain Behav. 2007;6:54–65. doi: 10.1111/j.1601-183X.2006.00221.x. [DOI] [PubMed] [Google Scholar]
  31. Xu G, Karch C, Li N, Lin N, Fromholt D, Gonzales V, et al. Receptor-associated protein (RAP) plays a central role in modulating Aβ deposition in APP/PS1 transgenic mice. PLoS ONE. 2008;3:e3159. doi: 10.1371/journal.pone.0003159. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Han BH, Zhou ML, Abousaleh F, Brendza RP, Dietrich HH, Koenigsknecht-Talboo J, et al. Cerebrovascular dysfunction in amyloid precursor protein transgenic mice: contribution of soluble and insoluble amyloid-β peptide, partial restoration via γ-secretase inhibition. J Neurosci. 2008;28:13542–13550. doi: 10.1523/JNEUROSCI.4686-08.2008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Jiang Q, Lee CY, Mandrekar S, Wilkinson B, Cramer P, Zelcer N, et al. ApoE promotes the proteolytic degradation of Aβ. Neuron. 2008;58:681–693. doi: 10.1016/j.neuron.2008.04.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. LaFerla FM, Green KN., and , Oddo S. Intracellular amyloid-β in Alzheimer's disease. Nat Rev Neurosci. 2007;8:499–509. doi: 10.1038/nrn2168. [DOI] [PubMed] [Google Scholar]
  35. Ohyagi Y. Intracellular amyloid β-protein as a therapeutic target for treating Alzheimer's disease. Curr Alzheimer Res. 2008;5:555–561. doi: 10.2174/156720508786898514. [DOI] [PubMed] [Google Scholar]
  36. Schwarz DS, Ding H, Kennington L, Moore JT, Schelter J, Burchard J, et al. Designing siRNA that distinguish between genes that differ by a single nucleotide. PLoS Genet. 2006;2:e140. doi: 10.1371/journal.pgen.0020140. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Grimm D, Streetz KL, Jopling CL, Storm TA, Pandey K, Davis CR, et al. Fatality in mice due to oversaturation of cellular microRNA/short hairpin RNA pathways. Nature. 2006;441:537–541. doi: 10.1038/nature04791. [DOI] [PubMed] [Google Scholar]
  38. Boudreau RL, Martins I., and , Davidson BL. Artificial microRNAs as siRNA shuttles: improved safety as compared to shRNAs in vitro and in vivo. Mol Ther. 2009;17:169–175. doi: 10.1038/mt.2008.231. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. McBride JL, Boudreau RL, Harper SQ, Staber PD, Monteys AM, Martins I, et al. Artificial miRNAs mitigate shRNA-mediated toxicity in the brain: implications for the therapeutic development of RNAi. Proc Natl Acad Sci USA. 2008;105:5868–5873. doi: 10.1073/pnas.0801775105. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Kuwabara T, Warashina M, Sano M, Tang H, Wong-Staal F, Munekata E, et al. Recognition of engineered tRNAs with an extended 3' end by Exportin-t (Xpo-t) and transport of tRNA-attached ribozymes to the cytoplasm in somatic cells. Biomacromolecules. 2001;2:1229–1242. doi: 10.1021/bm0101062. [DOI] [PubMed] [Google Scholar]
  41. Kuwabara T, Warashina M, Koseki S, Sano M, Ohkawa J, Nakayama K, et al. Significantly higher activity of a cytoplasmic hammerhead ribozyme than a corresponding nuclear counterpart: engineered tRNAs with an extended 3' end can be exported efficiently and specifically to the cytoplasm in mammalian cells. Nucleic Acids Res. 2001;29:2780–2788. doi: 10.1093/nar/29.13.2780. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Kawasaki H., and , Taira K. Short hairpin type of dsRNAs that are controlled by tRNA(Val) promoter significantly induce RNAi-mediated gene silencing in the cytoplasm of human cells. Nucleic Acids Res. 2003;31:700–707. doi: 10.1093/nar/gkg158. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Nussbaum RL., and , Ellis CE. Alzheimer's disease and Parkinson's disease. N Engl J Med. 2003;348:1356–1364. doi: 10.1056/NEJM2003ra020003. [DOI] [PubMed] [Google Scholar]
  44. Dreses-Werringloer U, Lambert JC, Vingtdeux V, Zhao H, Vais H, Siebert A, et al. A polymorphism in CALHM1 influences Ca2+ homeostasis, Aβ levels, and Alzheimer's disease risk. Cell. 2008;133:1149–1161. doi: 10.1016/j.cell.2008.05.048. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Urabe M, Ding C., and , Kotin RM. Insect cells as a factory to produce adeno-associated virus type 2 vectors. Hum Gene Ther. 2002;13:1935–1943. doi: 10.1089/10430340260355347. [DOI] [PubMed] [Google Scholar]
  46. Miller VM, Nelson RF, Gouvion CM, Williams A, Rodriguez-Lebron E, Harper SQ, et al. CHIP suppresses polyglutamine aggregation and toxicity in vitro and in vivo. J Neurosci. 2005;25:9152–9161. doi: 10.1523/JNEUROSCI.3001-05.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Harrison FE, Reiserer RS, Tomarken AJ., and , McDonald MP. Spatial and nonspatial escape strategies in the Barnes maze. Learn Mem. 2006;13:809–819. doi: 10.1101/lm.334306. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Patil SS, Sunyer B, Höger H., and , Lubec G. Evaluation of spatial memory of C57BL/6J and CD1 mice in the Barnes maze, the Multiple T-maze and in the Morris water maze. Behav Brain Res. 2009;198:58–68. doi: 10.1016/j.bbr.2008.10.029. [DOI] [PubMed] [Google Scholar]
  49. Hammond RS, Tull LE., and , Stackman RW. On the delay-dependent involvement of the hippocampus in object recognition memory. Neurobiol Learn Mem. 2004;82:26–34. doi: 10.1016/j.nlm.2004.03.005. [DOI] [PubMed] [Google Scholar]
  50. Scholtzova H, Wadghiri YZ, Douadi M, Sigurdsson EM, Li YS, Quartermain D, et al. Memantine leads to behavioral improvement and amyloid reduction in Alzheimer's-disease-model transgenic mice shown as by micromagnetic resonance imaging. J Neurosci Res. 2008;86:2784–2791. doi: 10.1002/jnr.21713. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Figure S1.

Design of allele-specific RNAi. Our RNAi targeting strategy placed the two nucleotide mismatch between wild-type APP and mutant APPsw alleles (GA-->UC) at the 10th and 11th nucleotide positions of the predicted shRNA guide strand (light gray font). Uninterrupted complimentary base pairing between APPsw mRNA (left column) and the APPsw shRNA guide strand results in RISC-dependent cleavage and degradation of target RNA. The double purine-purine mismatch between wild-type APP mRNA (right column) and the guide strand targeting mutant APPsw is predicted to disrupt RISC-dependent activity on wild-type APP mRNA. This disruption, likely due to the chemical “bulkiness” inherent to purine-purine interactions, spares the wild-type APP allele.

Click here for supplemental data (8.7MB, tiff)
Figure S2.

Gross morphological analysis of AAV-mediated RNAi transduction. Coronal sections obtained from wild-type and AD transgenic mice injected with either rAAV-shAPPsw or rAAV-shMiss were analyzed for gross morphological abnormalities following short-term rAAV-RNAi expression. (a) Hematoxylin and eosin (H&E) staining of mouse hippocampus. Age-matched comparisons between injected (AAV shMiss and AAV shAPPsw) and uninjected control AD transgenic mice did not reveal gross differences in the number of neurons or hippocampal structure. (b) Analysis of NeuN expression throughout transduced hippocampus confirmed results obtained in a. Scale bar= 450μm.

Click here for supplemental data (42.9MB, tiff)
Figure S3.

Efficient widespread silencing of APPsw expression. (a–h) Five weeks post-rAAV delivery, coronal sections were analyzed for both hrGFP and APPsw expression. rAAV-RNAi-mediated (shAPPsw and shMiss) transduction was most evident throughout the CA1 and CA2 regions of the anterior hippocampus as shown by hrGFP expression (a–d). Loss of APPsw expression (6E10 antibody) was most pronounced in areas of the hippocampus transduced with rAAV-shAPPsw virus (e, f). In contrast, contralateral expression of rAAV-shMiss had no effect on the expression pattern or intensity of APPsw transprotein in the contralateral hippocampus (g, h). Images presented are from two different mice and are representative of all other mice analyzed (n = 5). (i–n) To confirm this result, sections were subjected to immunohistochemical analysis using 3-3'diaminobenzidine as a color substrate. Again, we observed reduced levels of APPsw immunoreactivity throughout transduced regions of the hippocampus (i, j), compared to contralateral hippocampus expressing negative control virus, rAAV-shMiss (l, m). (k) Digital zoom of j showing the transduction boundary with a clearly demarcated difference in APPsw expression. (n) Digital zoom of m showing the presence and continuity of APPsw expression throughout the rAAV-shMiss transduced hippocampus. Scale bar = 250μm.

Click here for supplemental data (44.9MB, tiff)
Figure S4.

Age-dependent increase in amyloid burden in AD transgenic mice (APPsw/PS1Δ9). Congo red dye was used to visualize plaque burden in the brains of AD transgenic mice that were 4, 6, 8 or 12 months of age. Congo red-positive plaques begin to appear in the hippocampus of AD mice by 6 months of age (middle panels) and continue to increase in size and number at 8 months of age (right panels). Four month-old transgenic mice did not have Congo red-positive deposits whereas 12 month-old transgenic presented with the most amyloid load of all ages tested (not shown). Wild-type littermate controls were devoid of Congo red staining. WT = wild-type littermates, Tg = AD transgenic, BF= bright field, CR= Congo red signal. Scale bar = 250μm.

Click here for supplemental data (10.1MB, tiff)
Figure S5.

Suppression of APPsw levels in bilaterally rAAV-shAPPsw injected AD transgenic mice (APPsw/PS1Δ9). Total protein was extracted from transduced hippocampi and analyzed for transgene expression using the 6E10 antibody as described in figure 4. Long term expression of rAAV-shAPPsw in the hippocampus of AD transgenic mice (lanes 2 and 3) led to a significant reduction in the levels of APPsw protein when compared to rAAV-shMiss control injected (lanes 5 and 6) or uninjected age-matched AD mice (lanes 1 and 4). Tubulin levels were used as loading control (bottom). The number under the group heading (rAAV-shAPPsw or rAAV-shMiss) identifies individual subjects within each group.

Click here for supplemental data (9.5MB, tiff)

Articles from Molecular Therapy: the Journal of the American Society of Gene Therapy are provided here courtesy of The American Society of Gene & Cell Therapy

RESOURCES