Constitutive expression of short hairpin RNA in vivo triggers buildup of mature hairpin molecules

Abstract

RNA interference (RNAi) has become the cornerstone technology for studying gene function in mammalian cells. In addition, it is a promising therapeutic treatment for multiple human diseases. Virus-mediated constitutive expression of short hairpin RNA (shRNA) has the potential to provide a permanent source of silencing molecules to tissues, and it is being devised as a strategy for the treatment of liver conditions such as hepatitis B and hepatitis C virus infection. Unintended interaction between silencing molecules and cellular components, leading to toxic effects, has been described in vitro. Despite the enormous interest in using the RNAi technology for in vivo applications, little is known about the safety of constitutively expressing shRNA for multiple weeks. Here we report the effects of in vivo shRNA expression, using helper-dependent adenoviral vectors. We show that gene-specific knockdown is maintained for at least 6 weeks after injection of 1 × 10(11) viral particles. Nonetheless, accumulation of mature shRNA molecules was observed up to weeks 3 and 4, and then declined gradually, suggesting the buildup of mature shRNA molecules induced cell death with concomitant loss of viral DNA and shRNA expression. No evidence of well-characterized innate immunity activation (such as interferon production) or saturation of the exportin-5 pathway was observed. Overall, our data suggest constitutive expression of shRNA results in accumulation of mature shRNA molecules, inducing cellular toxicity at late time points, despite the presence of gene silencing.

FIG. 1.

Gene silencing is maintained over time. Mice were given 1 × 10¹¹ VP of gAd.shFABP5 or gAd.shSCR intravenously (n = 3). Animals were killed 1 or 8 weeks later. (A) Levels of FABP5 protein were quantified by Western blot. Lanes correspond to individual animals. (B) Densitometry analysis of FABP5 levels. Data represents averaged results of individual animals.

FIG. 2.

Long-term shRNA expression results in loss of vector genomes. Mice were given 1 × 10¹¹ VP of gAd.shSREBP1, gAd.shSCR, or gAd.NEC, or vehicle, intravenously and were killed 1, 3, and 6 weeks postinjection (n = 3). (A) SREBP1 gene silencing in the liver was evaluated by Western blot analysis. Lanes correspond to individual animals. Cyp, cyclophilin. (B) DNA was isolated from liver and vector genome copy number was analyzed by real-time PCR. DNA vector from the shRNA-expressing vectors decreased between weeks 3 and 6. (C) ALT levels were increased over time only in shRNA-expressing groups, not in vehicle and NEC (no expression cassette) groups. (D) The microRNA fraction was extracted from liver and shRNA expression was quantified by Northern blot. The level of shRNA expression was highest at 3 weeks and then declined, correlating with the level of vector DNA. Numbers 1 to 3 at the top of the lanes represent mice that received adenoviral vector. Lane 4 represents a mouse treated with vehicle (negative control for shRNA expression). (E) Densitometric analysis of shRNA expression relative to 5S.

FIG. 3.

shRNA expression elicits buildup of molecules over time. Mice were administered 1 × 10¹¹ VP of gAd.shSREBP1, gAd.shSCR, or gAd.NEC, or vehicle, and were killed weekly for 5 weeks (n = 3). (A) Body weight was not altered significantly among treatment groups. (B) shRNA expression was quantified by Northern blot (left) and densitometric (right) analysis of shRNA levels. Lanes correspond to individual animals. (C) ALT levels. (D) Northern blot of the microRNA-enriched fraction, using oligonucleotide (oligo) sequences to detect shSCR or shSREBP molecules. An RNA oligo with the precursor (50-mer) or mature (21-mer) SREBP sequence was spiked into the RNA of a vehicle-treated animal to determine the size of the shRNA molecules present in liver. Only molecules corresponding to the processed shRNA were detected. DNA SCR, scrambled DNA 21-nucleotide antisense sequence; DNA SREBP, SREBP DNA 21-nucleotide antisense sequence; RNA 21-mer, SREBP mature 21-nucleotide RNA oligo; RNA 50-mer, SREBP precursor 50-nucleotide oligo. (E) miR-122 levels were not altered at any time points.

FIG. 4.

Interferon-stimulated gene (ISG) expression. Mice were treated as described in Fig. 3. Expression levels of two ISG genes, ISG56 and Oas1b, were quantified by real-time RT-PCR. No increase was observed in the gAd.shSREBP and gAd.shSCR groups compared with the gAd.NEC group.

FIG. 5.

Eri-1 gene and protein analysis. Mice were treated as described in Fig. 3. (A) mRNA expression analysis of Eri-1 in liver by real-time RT-PCR. (B) Western blot quantification of Eri-1 protein. No differences were observed in any group at any time point.

FIG. 6.

Histology of mouse liver sections. Mice were treated as described in Fig. 3. Liver sections were stained with hematoxylin–eosin and analyzed by a liver pathologist. Panels display the central vein area. In gAd.shSCR-treated liver, dense cells suggestive of having cellular toxicity are denoted by arrows, and normal cells are identified by arrowheads. In animals that received the gAd.shSREBP vector, dense cells were observed almost everywhere in the liver. Arrows indicate the presence of oval cells.