Targeting Alzheimer's disease genes with RNA interference: an efficient strategy for silencing mutant alleles

Abstract

Tau and amyloid precursor protein (APP) are key proteins in the pathogenesis of sporadic and inherited Alzheimer's disease. Thus, developing ways to inhibit production of these proteins is of great research and therapeutic interest. The selective silencing of mutant alleles, moreover, represents an attractive strategy for treating inherited dementias and other dominantly inherited disorders. Here, using tau and APP as model targets, we describe an efficient method for producing small interfering RNA (siRNA) against essentially any targeted region of a gene. We then use this approach to develop siRNAs that display optimal allele-specific silencing against a well-characterized tau mutation (V337M) and the most widely studied APP mutation (APPsw). The allele-specific RNA duplexes identified by this method then served as templates for constructing short hairpin RNA (shRNA) plasmids that successfully silenced mutant tau or APP alleles. These plasmids should prove useful in experimental and therapeutic studies of Alzheimer's disease. Our results suggest guiding principles for the production of allele-specific siRNA, and the general method described here should facilitate the production of gene-specific siRNAs.

Figure 1

siRNA+G duplexes silence endogenous and reporter genes. (A) Schematic of siRNA synthesis depicting DNA template and structure of synthesized duplexes. Blue indicates the RNA product synthesized from the DNA template (upper). For the siRNA duplex, gray indicates the region with perfect complementarity to the intended target while black depicts the sense sequence and additional non-complementary nucleotides added by the synthesis method. N represents any ribonucleotide. (B) Comparison of GFP silencing by perfectly complementary siRNA versus siRNA of the ‘+G’ design. Images depict Cos-7 cells transfected with a GFP expression construct and the indicated siRNA. Images of GFP fluorescence are merged with images of the same field showing DAPI-stained nuclei. Shown on the left are results with negative control, mistargeted siRNAs (siMiss and siMiss+G respectively), which fail to silence GFP expression. On the right, GFP expresssion is efficiently suppressed by siRNA of both configurations. (C) Western blot analysis of lysates from the same experiment as in (B). Tubulin staining is shown as a loading control. (D) Efficient silencing of endogenous lamin gene expression with siRNA+G duplexes. HeLa cells were transfected with the indicated siRNA and expression of lamin A/C was evaluated by western blot 72 h later. The siRNA+G against human lamin markedly decreased protein levels relative to the mistargeted control siRNA. All western blots shown are representative of two independent experiments.

Figure 2

Optimization of allele-specific silencing of mutant tau. Cos-7 cells were cotransfected with expression constructs encoding mutant (V337M-GFP) and WT (Flag-WT) tau and the indicated siRNAs or shRNA plasmids. (A) Western blot results showing the efficacy of allele-specific silencing when varying the placement of the point mutation (G to A) in the siRNA from positions 9–12. (B) Western blot analysis of cells cotransfected as in (A). Amounts of expression plasmids were held constant while the concentration of siA10 was varied from 0 to 10 µg per well. (C) Silencing tau with shRNA plasmid expressed from the tRNA-valine promoter. Shown is a western blot analysis of cells cotransfected with mutant and wild type tau and the indicated shRNA plasmids. Placing the mutation at position 10 (tvA10) of the hairpin results in strong preferential silencing of mutant tau. shRNA directed against wild type (mismatched at position 9 relative to mutant tau) tau inhibits expression from both alleles but shows a preference for the wild type sequence. (D) Quantitation of western blot signal from two independent experiments performed as in (C). Bars depict mean signal intensity and standard deviations between experiments. Signal from cells co-transfected with tvMiss was set at one. All western blots shown are representative of two independent experiments.

Figure 3

Optimization of allele-specific silencing of mutant APP. Cos-7 cells were transfected with expression constructs encoding wild type APP (APP) or mutant (APPsw) and the indicated siRNAs or shRNA plasmids. Silencing was evaluated in separate transfections of the wild type and mutant constructs because the 22C11 antibody detects both APP and APPsw proteins. (A) Immunofluorescence of Cos-7 cells cotransfected with plasmids encoding APP or APPsw and the indicated siRNA. Representative images of fields (630×) reveals that allele specificity is optimal when the double mismatch is placed at the central position (siT10/C11) of the targeted sequence. APP proteins are visualized with APP antibody followed by secondary antibody labeled with FITC (green). Nuclei are stained with DAPI (blue). (B) Lanes 5–10 show a western blot of cells transfected as in (A), confirming preferential silencing of APPsw with siRNA containing central mismatches. Lane 4 is APP or APPsw transfected without siRNA. Lane 11 represents untransfected cells showing endogenous APP. Also shown in lanes 1–3 is comparable silencing of APP with siRNA or siRNA+G duplexes targeted to APP. Tubulin is shown as a loading control. (C) Western blot analysis of Cos-7 cells transfected with APP or APPsw and the indicated shRNA plasmids. tvAPP silences APP whereas tvT10/C11 selectively suppresses APPsw expression. Endogenous APP in untransfected cells is shown in the last lane. Tubulin loading control is also shown. Western blots and immunofluorescence images are representative of two (C and D) or four (A and B) independent experiments. (D) Quantitation of two independent western blot experiments performed as in (C). Bars depict mean signal intensity and standard deviations between experiments. Signal from cells co-transfected with tvMiss was set at 1.