Skip to Main Content

Cover Story

November 13, 2006
Volume 84, Number 46
pp. 16-23

Delivering RNA Interference

Developing siRNA therapeutics depends on synthetic delivery systems

Celia Henry Arnaud

RNA interference is on the fast track. In the eight brief years since the RNAi gene-silencing mechanism was first uncovered, its discoverers have won the Nobel Prize and the first therapeutics based on it have entered clinical trials. The announcement just two weeks ago that Merck will acquire San Francisco-based Sirna Therapeutics, which is one of the major players in RNAi-based therapeutics, for $1.1 billion shows that big pharma is confident about the potential of RNAi therapeutics (C&EN, Nov. 6, page 11).

Target Purple-and-green conjugates of peptides and double-stranded RNA target cell-surface receptors (purple) and deliver double-stranded RNA to the Dicer enzyme (orange), which cuts the RNA to the right size for RNA interference.

Therapeutics based on RNAi take advantage of this natural gene-silencing mechanism. They take the form of small, double-stranded RNA molecules just 19 to 21 nucleotides long—so-called small interfering RNAs, or siRNAs—which guide complementary messenger RNA to a protein complex known as RISC. RISC then cleaves the mRNA and prevents its translation into protein.

In the first group of RNAi therapeutics, the siRNAs are administered directly to the disease location. For example, Sirna's lead candidate is directly injected into the eye to treat age-related macular degeneration, and Cambridge, Mass.-based Alnylam Pharmaceuticals' treatment for respiratory syncytial virus is delivered directly to the lung by inhalation. Sirna's candidate, which is being developed in partnership with Allergan, is in Phase II clinical trials, and Alnylam has recently launched the third Phase I clinical trial of its candidate.

"There are a number of diseases where local delivery would be all you need," says Judy Lieberman, a researcher at the CBR Institute for Biomedical Research at Harvard Medical School. But diseases that can be treated by such local administration are ultimately limited, and "systemic delivery is still a problem," Lieberman says.

For RNAi to have the therapeutic impact that many people hope it will have, systemic delivery methods are needed. Such delivery systems are the focus of intense investigation by industrial and academic researchers.

Although nucleic acid therapeutics have been around in various forms for approximately three decades, they haven't yet been successful in the clinic, according to Barry Polisky, senior vice president of research and chief scientific officer at Sirna. "Nucleic acid therapeutics has really been an idea whose promise has not yet been realized," he says. "It's almost entirely due to the lack of attention paid to the delivery problem. It's hard to emphasize enough the centrality of this issue."

The first examples of systemic delivery have been to the liver, for which multiple methods are proving feasible. Alnylam conjugated cholesterol to siRNA targeting the gene for apolipoprotein B and showed that systemic administration in mice resulted in less production of the apoB protein in the liver (Nature 2004, 432, 173). Alnylam also reduced apoB in monkeys by delivering the same siRNA using a lipid-based nanoparticle delivery system designed by Protiva Biotherapeutics, Burnaby, British Columbia (Nature 2006, 441, 111).

"As optimistic as we all are about delivery, there's a lot of hard work left to do if we want to deliver outside of the liver," says Phillip D. Zamore, who studies RNAi at the University of Massachusetts Medical School, in Worcester. Still, "delivery to the lung looks promising," Zamore points out. And even if there is more work to do to target other organs, "if you can deliver to lung and liver, there's plenty of human suffering you want to alleviate in those tissues."

John J. Rossi, a molecular biologist at the Beckman Research Institute at City of Hope in Duarte, Calif., thinks the problems are closer to being solved. "A number of good strategies have been published in the last year and a half that suggest we have a bunch of choices now" for systemic siRNA delivery, he says, including lipid nanoparticles, cyclodextrin-based polymers, and RNA ligands called aptamers.

"We do not believe there is going to be one universal delivery solution," says Nagesh Mahanthappa, senior director of business development and strategy at Alnylam. Instead, it will be important to create a "palette of technologies" from which to choose on the basis of the disease and cell type.

Although systemic delivery is universally accepted as key to RNAi therapeutics, one aspect of it is still being debated, and that is whether siRNAs should be chemically modified when delivered.

Sirna Therapeutics
Analysis Weimin Wang performs NMR studies to understand the pH-dependent structural changes in Sirna's lipid nanoparticles.

Chemical modifications are essential when therapeutic siRNAs are introduced without a delivery vehicle. Such modifications are intended to boost the siRNA's stability in the blood by protecting the RNA from enzymes known as nucleases, which chew up nucleic acids, and to prevent the siRNAs from triggering an immune response. But experts disagree whether such modifications are necessary when the siRNA is protected by a delivery system.

"Any time you do a chemical modification of the siRNA, that's not RNA any more," says Mark E. Davis, a chemical engineering professor at California Institute of Technology who is working on a cyclodextrin-based siRNA delivery system. The modified siRNA degrades into molecules that aren't naturally found in the body, making the decomposition products an additional safety concern, he notes.

Properly designed delivery systems can mask the unmodified siRNA so that it can reach the cell without causing immune responses, he says. "We have performed careful studies over the past two years to confirm the surprising observations of several groups that non-chemically modified siRNAs can provide gene inhibition" that lasts as long as that due to chemically modified siRNAs, Davis says.

According to Steven C. Quay, chairman, president, and chief executive officer of Nastech Pharmaceutical, in Bothell, Wash., siRNAs do not have to be any more stable than is required for them to reach their target cells. "If you have an effective delivery system, so that you get delivery when the material encounters the proper cell, you don't really need to have hours and hours of stability in the blood stream," he says. "To my mind, a delivery system that creates stability in the bloodstream of 24 hours simply means that it doesn't get into cells."

Lieberman thinks that modified RNA may even be a disadvantage inside the cell. "The natural machinery was developed for unmodified siRNAs," she says. "Once you modify them, you're going to interfere with the efficiency. However, there may be some small modifications that will buy you better specificity and reduce certain kinds of toxicity."

Researchers at Sirna see things differently. Work at that company demonstrates the "critical need for modification" even with efficient delivery, Polisky says. "The cell contains very potent nucleases that can degrade these RNAs inside the cell," he explains. Another difference between modified and unmodified RNA is in duration of effect, he says. Sirna compared the duration of effect of modified and unmodified siRNAs delivered with lipid nanoparticles. "We saw very dramatic differences in performance, where the modified RNA was very superior."

In addition, Polisky says, the modified siRNAs may avoid triggering an immune response. He explains that double-stranded RNA can elicit an immune response that involves the secretion of chemicals known as cytokines. "If we modify the RNA the way we have traditionally modified it, we actually can suppress this phenomenon very dramatically," he says. "The cell doesn't really sense the presence of these chemically modified double-stranded RNAs as an alarm signal." Polisky believes that the cytokine response may be largely responsible for off-target effects seen with siRNA.

Even though the debate on chemical modification is not yet settled, researchers at companies and in academia are working on a variety of delivery systems. The most developed are lipid-based nanoparticles.

Alnylam and Protiva used Protiva's SNALP (stable nucleic acid lipid particles) technology to deliver siRNA targeting the apoB gene in monkeys. These lipid particles consist of cationic lipids, lipids that can fuse with cell membranes (so-called fusogenic lipids), lipids conjugated to polyethylene glycol (PEG), and the siRNA. The ratios of the lipid components can be varied to change the cell type that takes up the particle. The length of the PEG-conjugated lipid affects the circulating half-life and tissue distribution of the particles. Rather than loading siRNA into preformed delivery vehicles, the particle is assembled around the siRNA.

Alnylam Pharmaceuticals
Snip, Snip siRNA, shown here conjugated to a targeting molecule such as cholesterol, enters the RISC complex and guides mRNA to be cleaved, leading to mRNA degradation and gene silencing.

"We were the first group to demonstrate that one could administer an siRNA in a nonhuman primate and see silencing of a target gene," says Alnylam's Mahanthappa. "In this particular study, the liposomes were optimized for uptake by liver cells, but I think liposomes will prove to be a broadly applicable technology."

Sirna is also working on a lipid nanoparticle that encapsulates the siRNA. "These lipids are designed to change under certain biological conditions," says Chandra Vargeese, vice president of delivery at Sirna. The lipid nanoparticles are taken up by the cells via endocytosis, a process by which materials are brought into cells inside acidic vesicles known as endosomes. Once inside the endosomes, the nanoparticles' lipids undergo pH-dependent changes that disrupt the nanoparticles and release the siRNA.

Sirna has started to engineer lipid nanoparticles that target tissues other than the liver. The system is "in advanced stages for certain targets like the liver, and it's in early stages for other targets," Polisky says.

Yet another company developing a lipid nanoparticle-based delivery system is Berlin, Germany-based Atugen. The company's Atuplex system consists of a cationic lipid and a fusogenic lipid that can disrupt the endosomal membrane. The company has generated a panel of lipid nanoparticle formulations that target different cell types.

Not all nanoparticles used for siRNA delivery are lipid based. The cyclodextrin nanoparticle delivery system of Calando Pharmaceuticals, Pasadena, Calif., uses cyclodextrin-based polymers developed by Davis, one of the company's founders. In the delivery system, a cyclodextrin-containing polycation is mixed with a conjugate of adamantane and PEG and a separate three-part conjugate of adamantane, PEG, and a targeting ligand. The adamantane forms an inclusion complex with the cyclodextrins within the polycation, allowing noncovalent incorporation of the PEG-containing components within the complex.

"We prepare our formulations by premixing the three delivery components in one vial and adding this mixture to a solution of siRNA in another vial," says Jeremy D. Heidel, vice president of research and development and chief scientific officer at Calando. "This formulation strategy gives us the potential to 'mix and match' various targeting ligands and siRNAs." Heidel presented data at the Oligonucleotide Therapeutics Society (OTS) meeting in New York City last month showing that that multiple doses of Calando's formulation, which employs unmodified siRNA, could be safely administered to monkeys and not cause immune responses.

Courtesy of Derek Bartlett/Caltech
Bumpy Cyclodextrin-containing polymers form complexes with siRNA that appear as bumps in this AFM image.

Another class of delivery system involves simply conjugating the siRNA to another molecule, such as cholesterol or peptides or even another RNA molecule.

One such delivery system is a peptide-based method developed by Nastech. The peptides are conjugated to pieces of double-stranded RNA that are 25 to 30 nucleotides long, which is slightly longer than typical siRNAs. These longer nucleotides are substrates for the Dicer enzyme, which, as its name implies, is responsible for cutting double-stranded RNA into the short pieces that work with the RISC complex. Thus, after delivery, the peptide gets cut away in the process of liberating the siRNA that will knock the targeted gene down.

Nastech has two RNAi therapeutic programs, both of which use this delivery method. One targets tumor necrosis factor α in rheumatoid arthritis, and the other targets genes in the influenza virus.

Meanwhile, two academic groups have recently shown that RNA ligands called aptamers can be used to guide siRNAs to their targets. Andrew D. Ellington of the University of Texas, Austin, and Bruce A. Sullenger of Duke University have independently demonstrated the use of aptamers to target siRNA to prostate cancer cells.

The groups chose different previously identified aptamers that target the same receptor found on prostate cancer cells. Ellington connected the aptamer to the siRNA via a biotin-streptavidin link (Nucl. Acids Res., DOI: 10.1093/nar/gkl388). Sullenger's group connected the two RNA molecules via a double-stranded RNA linker (Nat. Biotechnol. 2006, 24, 1005).

Uptake Lipid nanoparticles (red) decorate a mouse liver vein. Cell nuclei are stained green.

Ellington has thus far demonstrated the aptamer-mediated delivery only in cell culture, but Sullenger's group has progressed to a mouse model of prostate cancer. Speaking at the OTS meeting, Sullenger said that although the aptamer delivery is not yet quite as good as lipid delivery, significant reduction in the translation of the targeted gene is still possible.

In addition to protecting the siRNA, delivery systems can also escort the siRNA to specific cells. The importance of such targeted delivery varies with the gene of interest. In some cases it's a necessity and in others simply a bonus, but it's probably desirable in all cases.

"Cell-specific delivery provides enormous advantages, both in terms of the concentration or the dose of siRNA you need to get a therapeutic effect as well as the likely reduction in possible side effects and toxicity," Lieberman says.

If the gene is expressed only in certain cells, the siRNA will be effective only in those cells. In that case, side effects are probably not an issue. For genes that are expressed in many cells but are a problem only in a subset of those cells, the delivery method should home in on those cells to avoid knocking the gene down elsewhere.

A number of different types of targeting ligands are being investigated. Nastech's peptides serve as targeting ligands. Likewise, Calando is currently using the transferrin protein with its cyclodextrin delivery system to target receptors on the surface of cancer cells.

Lieberman is examining the use of fusion proteins made of antibodies and protamine, a protein that condenses nucleic acids, as simultaneous delivery and targeting agents. Because the fusion protein forms a noncovalent complex with the siRNA to be delivered, the same reagent can be used to deliver any siRNA. Such constructs have been shown to deliver siRNA to cells infected with HIV (Nat. Biotechnol. 2005, 23, 709). Lieberman is currently developing reagents to target different blood cells and immune cells. She has licensed the technology to Alnylam.

Although most siRNA delivery solutions are chemical in nature, "there are also engineering solutions," Mahanthappa says. "A good example of that is our partnership with Medtronic. We're exploring the use of Alnylam's RNAi technology with Medtronic's medical devices to deliver siRNAs directly into the central nervous system." The partnership is focused on treatments for neurodegenerative diseases such as Huntington's and Parkinson's diseases.

But delivery methods are pointless without therapeutic targets. Some of the initial therapeutic targets of siRNA are infectious diseases, particularly those caused by viruses. "Viruses are a good choice for RNAi drug development," Lieberman says. A virus offers gene targets that are specific only to the viral disease, and silencing a viral gene affects only the virus. By contrast, treating metabolic diseases by targeting certain human genes could have the unintended effect of silencing genes that are necessary for normal health. "There are very few viruses for which we have really good drugs," she adds.

At the OTS meeting, Ian MacLachlan, chief scientific officer of Protiva, reported his company's progress in developing an siRNA treatment for Ebola virus. The system, currently being tested in a guinea pig model, targets the viral gene for an enzyme known as L-polymerase. To treat Ebola at the earliest stage of the infection, the Protiva scientists are targeting dendritic immune cells in the blood rather than hepatocytes in the liver, but to treat later stages of the disease, they will also need to reach liver cells. "The formulation is a compromise between accessing blood cells and liver cells," MacLachlan said.

Sirna is also working on targeting a virus, namely hepatitis C. Because of the lack of an animal model of hepatitis C, the company is currently working with a hepatitis B mouse model, according to David V. Morrissey, senior director of antiviral therapeutics. Speaking at the OTS meeting, he reported that Sirna is getting a sustained knockdown of the hepatitis B virus in these mice from a dosing regimen of two doses in the first week, followed by one dose a week after that. Sirna's hepatitis C clinical candidate uses a mixture of two siRNAs, and it targets sites that are found in more than 96% of clinical hepatitis C samples.

Next year could see the first clinical trials of RNAi therapeutics packaged in delivery systems. Calando hopes to start Phase I clinical trials for cancer by the end of 2007. Similarly, Sirna and Atugen both plan clinical trials of therapeutics based on their nanoparticle delivery systems, Sirna for hepatitis C and Atugen for cancer.

"Next year could be an interesting year in the sense that actual delivered siRNA could be moving into the clinic," Davis says. "It's nice to see all the data in animals, but it's a huge jump to move to the clinic with these synthetic systems. Someone getting to the clinic, that's a whole new ballgame."

Cover Story

Cover Story

Related Stories

Adjust text size:

A- A+

Articles By Topic