RNAi Applications: Examples from the Literature

Predictions about mammalian gene function are sometimes based on methods for identifying differentially expressed genes or on the known function of homologous genes in model organisms like Drosophila, C. elegans, and S. cerevisiae. RNA interference allows scientists to knock down expression of a targeted gene and observe the effects of partial to full loss of function. 

Engagement of the interleukin-1 (IL-1) receptor and the Toll-like receptor (TLR) induces mitogen-activated protein kinase (MAPK) pathways and activation of transcription factors such as NFkB and AP-1. Three members of the novel Pellino protein family have been shown to be mediators in these pathways, but show pathway specificity in their function. Using pre-designed siRNA targeting human Pellino3, RNA interference was used to show a physiological role for Pellino3 in mediating IL-1 activation of p38 MAPK in HEK293 cells: Pellino3-specific siRNA caused a considerable reduction in IL-1 induced phosphorylation of p38 MAPK, relative to a control Pellino3-scrambled siRNA. The use of RNAi was validated by showing that total p38 MAPK protein levels were unaffected by either siRNA, Pellino3 mRNA levels were strongly suppressed by Pellino3-specific siRNA but not by the scrambled control, and GAPDH mRNA levels were unaffected by either siRNA.

Modifications of core histones, including phosphorylation, acetylation, methylation and ubiquitinylation, form the basis of a critical regulatory system in the control of chromatin structure. Haspin genes comprise a novel eukaryotic gene family that is found in all major eukaryotic phyla and encode proteins that contain a distinctive C-terminal putative kinase domain. In this study, the product of the human haspin gene was shown to phosphorylate histone H3 in vitro at Thr3, an amino acid that is highly conserved in histone H3 among all eukaryotes. Furthermore, histone H3 was shown to be phosphorylated at Thr3 during mitosis, and over-expression of haspin under an inducible promoter in HeLa cells caused increased histone H3 Thr3 phosphorylation and a delay in mitosis. The physiological role of haspin was shown through depletion of haspin via RNA interference, using an Ambion-validated siRNA targeting human haspin. Haspin siRNA-transfected U2OS cells showed reduced endogenous haspin kinase activity, dramatically reduced phosphorylation of histone H3 at Thr3 in mitosis, and misalignment of metaphase chromosomes, relative to control scrambled siRNA- or untransfected cells.

COX-2, an inducible form of cyclooxygenase, is frequently overexpressed in a variety of different malignancies. Previous studies indicate that COX-2 promotes cell survival by modulation of antiapoptotic protein expression. COX-2 over-expression leads to stabilization of survivin, a member of the inhibitor of apoptosis protein (IAP) family, in non-small cell lung cancer (NSCLC), resulting in elevated apoptosis resistance of COX-2-overexpressing cells. In this study, inhibition of survivin expression using siRNA targeting survivin in COX-2-overexpressing NSCLC cells reduced the threshold for apoptosis to that of parental NSCLC cell lines, while a negative control siRNA had no effect. The authors conclude that COX-2-dependent expression of survivin is critical for apoptosis resistance in NSCLC.

Cellular pathways that have been characterized biochemically or genetically can be further dissected using RNA interference. Libraries of siRNAs targeting large collections of genes enable screening experiments that tie genes to cellular function. If a cell-based assay that produces phenotypes relevant to a disease of interest is available, it can be used in high throughput screens using RNAi for identification of potential therapeutic targets. A key to a successful functional screening experiment is a robust phenotypic screening assay and a high quality siRNA library.

Pelkmans et al. used a library of siRNAs targeting a collection of 590 human protein, lipid, and carbohydrate kinases to screen for kinases involved in two principal types of endocytosis: clathrin-mediated and caveolae/raft-mediated. The screen used infection by vesicular stomatitis virus (VSV) or SV40, which hijack the clathrin-mediated, or caveolae-mediated and non-caveolar, lipid raft-mediated endocytic pathways, respectively, to infect host cells. They identified a large number of kinases involved in endocytosis and subsets of kinases involved in individual endocytic pathways, including one subset that exerted opposite effects on the monitored pathways. Phenotypic profiles of each silenced target were built based on assays for VSV or SV40 infection, transferrin trafficking, cell proliferation, and apoptosis. The kinases were then organized into functional groups through a two-step cluster analysis.

Caveolar vesicles are involved in cellular processes such as signal transduction and endocytosis. Using a combination of quantitative total internal reflection fluorescence microscopy (TIR-FM) and computational analysis, caveolae were classified into three major groups: caveolae which are not engaged in trafficking and are found as multi-caveolar assemblies that communicate with the extracellular environment; transport-competent caveolae undergoing constitutive kiss-and-run cycles within a small volume underneath the plasma membrane; and caveolae engaged in long-range cycling utilizing caveolae-mediated transport to other intracellular locations. Using these same assays, phenotypes reflecting function in caveolar coat stability, caveolae clustering, kiss-and-run dynamics, and long-range cycling were identified by RNAi, using siRNAs targeting six kinases identified in the previously described genomic-wide screen as potential regulators of caveolae/raft-mediated endocytosis.

A collection of siRNAs targeting 650 known and putative kinases was screened in HeLa cells using a histone-DNA fragmentation ELISA to provide a quantitative cell-based assay for apoptosis. These experiments identified 73 kinases (11% of the kinase library) whose knockdown resulted in increased apoptosis; the collection was termed “survival kinases.” Follow up experiments showed that the siRNA-induced apoptosis was blocked by caspase inhibitors zVAD or zDEVD; was characterized by increased amounts of activated caspase-9 and cleaved poly(ADP-ribose) polymerase (PARP), both hallmarks of apoptosis; and showed increased cell death as measured by crystal violet.

A similar screen for increased apoptosis was performed using an siRNA library against 222 known and putative phosphatases, using the DNA fragmentation ELISA and validated by production of cleaved caspase-9 and PARP. This screen identified 72 phosphatases (32% of the phosphatase library) involved in survival signaling, most of which were previously unidentified.  Several phosphatases that normally sensitize cells to cisplatin-induced apoptosis were also identified by transfecting cells with the human phosphatase siRNA library followed by treatment with cisplatin and assaying for reduced apoptosis by DNA fragmentation ELISA and increased cell viability using the crystal violet assay.

Several kinases were identified that, when down-regulated, enhance drug-induced apoptosis. HeLa cells were transfected with the siRNA kinase library in the absence or presence of low-dose concentrations of Taxol, etoposide, or cisplatin. Several kinases identified in this screen were shown to sensitize a breast carcinoma cell line to Taxol-induced apoptosis. These kinases represent potential targets for combination therapy with conventional chemotherapeutic agents.

Lysophosphatidic acid (LPA) and its G-protein-coupled receptors (GPCRs) LPA1, LPA2, or LPA3 have been implicated in the development of several types of cancers, including colorectal cancer. LPA was shown to stimulate cell proliferation in several colon cancer cell lines. RNAi using siRNAs targeting the individual receptors was used to show that LPA2 and LPA3, but not LPA1, were required for LPA-induced cell proliferation. RNAi was also used to show that LPA-induced proliferation is mediated by the b-catenin pathway, which is an important signaling pathway in the etiology of colon cancer.

In its simplest form, drug development follows the path of target identification -> target validation -> assay development -> drug lead identification -> lead optimization -> pre-clinical testing -> clinical trials. Once a potential therapeutic target has been identified through a variety of biochemical or genetic methods, including screening using RNAi, siRNAs can be used to reduce its expression. If the desired phenotype results, this outcome provides confidence that an inhibitor of the same target should have therapeutic value. In some cases, the siRNA itself may become the therapeutic lead.

Wet age-related macular degeneration (AMD) is characterized by neovascularization (overgrowth of new blood vessels under the retina) causing fluid leakage and scar tissue, leading to central vision impairment. The vascular endothelial growth factor (VEGF) pathway is thought to play a central role in retinal and choroidal neovascularization. Both VEGF and VEGF receptors (VEGFRs) are potential therapeutic targets. A panel of chemically stabilized siRNAs targeting vegfr1 (but not vegfr2) mRNA was screened in cultured endothelial cells for the ability to specifically knockdown expression of this gene, leading to identification of Sirna-027. Sirna-027 was shown to reduce levels of vegfr1 mRNA after intravitreous or periocular injection, compared to injection of an inverted control sequence. In a laser-induced model of choroidal neovascularization (CNV) in mice, intravitreous or periocular injections of Sirna-027 resulted in significant reductions in the area of ocular neovascularization (NV) ranging from 56% to 66%, compared to injection with inverted control siRNA or PBS in the fellow eye. In a mouse model of ischemia-induced retinal NV, Sirna-027 resulted in reduction of 32% in the area of retinal NV compared to injection of inverted control siRNA in the fellow eye.

In addition to their function in target validation studies, siRNAs have the potential to act as therapeutic products. Several studies have already been performed in mice and rats, and siRNAs are now known to be tolerated and effective in several different tissues. RNAi-based drugs currently in pre-clinical development include those targeting respiratory syncytial virus, hepatitis C, HIV, Huntington’s disease and several other neurodegenerative disorders [51, 52]. Phase I clinical trials are underway for several siRNA-based drugs targeting age-related macular degeneration (AMD), and Phase II trials have recently started for one siRNA candidate for AMD (www.sirna.com/sirnaproduct/sirna-027.html; www.acuitypharma.com/press.asp). Cellular delivery and stabilization are critical hurdles to overcome in development of viable siRNA-based drugs.

Reducing serum cholesterol and LDL levels is a primary clinical strategy for preventing and managing coronary artery disease. Apolipoprotein B (apoB) binds to the LDL receptor and is essential for formation of low-density lipoproteins (LDL) in cholesterol metabolism. Mice that are heterozygous for a knockout mutation of apoB have decreased cholesterol levels and are resistant to diet-induced hypercholesterolaemia. Conventional small-molecule or protein-based therapeutics targeting apoB have so far not been successfully developed.

84 siRNAs targeting mouse and human apoB in HepG2 liver cells were screened for their ability to reduce apoB mRNA and protein levels. Two of the most potent siRNAs were chemically modified and conjugated to cholesterol, which have been shown to confer improved pharmacological properties on siRNAs in vitro and in vivo. These siRNAs were shown to reduce apoB mRNA in liver and jejunum (the primary sites of apoB expression), decrease plasma levels of apoB protein, and reduce total cholesterol after intravenous injection in mice. Cleavage of apoB mRNA was shown to occur specifically at the site predicted by current models of RNAi, 10 nt downstream of the 5' end of the siRNA antisense strand.

Mutations in the Cu/Zn superoxide dismutase, encoded by SOD1, have been identified in approximately 20% of familial cases of amyotrophic lateral sclerosis (ALS). Transgenic mice overexpressing mutant forms of human SOD1 provide a model of ALS. A lentiviral vector encoding SOD1 shRNA (LV-shSOD1) targeting all forms of human SOD1 was shown in mouse embryonic fibroblasts derived from transgenic mice expressing human SOD1G93A to reduce mRNA and protein levels of SOD1G93A compared to a control construct containing two mismatch mutations in the shRNA stem sequence (LV-shSOD1mis). In vitro, LV-shSOD1 protected SOD1G93A motoneurons from nitric oxide-induced cell death. SOD1G93A mice injected in the lumbar spinal cord with LV-shSOD1 showed significantly delayed disease onset, delayed onset of the decline of neuromuscular function, and delayed progression rate of motor deficit and ameliorated motor strength, compared to mice injected with LV-shSOD1mis. LV-shSOD1G93A mice showed increased motoneuron survival near injection sites and significantly increased the number of large motor fibers compared to LV-shSOD1mis mice.