April 15, 2005 (Vol. 25, No. 8)
Ready for Prime Time
Post-transcriptional control though RNA interference (RNAi) is a remarkably well conserved mechanism found throughout eukaryotic organisms from plants through fungi, nematodes, and people.
RNAi, while it most likely evolved as one defense mechanism against marauding viral genomes and as a means of gene expression control during development, has also been very good for business.
RNAi has spawned a host of new enterprises, including those developing and applying RNAi-based technologies for drug discovery, functional genomic analysis, the development of new animal models for human diseases, and gene-silencing therapeutics to treat viral infections and cancer.
A phenomenon first discovered in plants and then in higher eukaryotes, RNAi functions as a naturally occurring mechanism that blocks expression of undesirable gene products through specific degradation of messenger RNAs.
RNAi is mediated by the enzymatic processing of long double-stranded RNA (dsRNA) into short-interfering RNA (siRNA)with sequence homology to the target mRNA. These siRNA molecules associate with a group of proteins forming a RNA-induced silencing complex, or RISC.
Once associated with RISC, the siRNA is unwound and directed toward the complementary target mRNA. RISC binds to the mRNA though base-pairing, specifically splicing the message and thereby preventing the subsequent protein expression.
RISC, still associated with the siRNA, moves on to the next target, silencing it in a catalytic fashion. This enzymatic mechanism differs from that of other drugs in which one drug molecule binds to one target molecule. Typically, about 5,000 mRNA copies exist following transcription of a particular gene.
Hailed as “powerful new tools” for determining gene function, and actively pursued as potential pharmaceuticals, siRNAs initially proved challenging to use. However since its initial application, siRNA technology has evolved to the point where scientists using it regard it as “more reliable and predictable” than antisense technology.
Transfection of siRNAs, however, induces complex responses in target cells, and interpretation of phenotypic alterations in comparison with small molecule drugs requires careful analysis. The success of companies pursuing therapeutics will depend on their ability to produce molecules with predictable pharmacological properties including biostability, pharmacokinetics, specificity, and the avoidance of so-called “off target” effects.
At an upcoming Cambridge Healthtech conference, “RNAi for Pathway Analysis,” scientists will talk about how their companies are applying RNAi technology in novel ways, as well as working to improve the performance and reliability of RNAi-related reagents and methods.
Small Molecule Drugs
Paul Young, Ph.D., vp of research at Avalon Pharmaceuticals (Germantown, MD), explains that his company is using siRNA-based tools to identify and validate targets for the company’s small molecule drugs. “We use RNAi to find otherwise inaccessible targets by combining it with molecular cytogenetics,” Dr. Young says.
“We have 170 unique amplified DNA regions in tumor DNA, each of which we believe contains a key gene that the cancer cell maintains because it drives the aggressive phenotype of the cancer cell.” The challenge is, he explains, that a lot of the targets are “not friendly” to small molecule screening.
For example, Dr. Young cites ATP or G-coupled receptors, regulatory pathways in which the key gene product may turn out to be transcription factors vital to normal cell function. Avalon uses a new screening approach blending expression profiling with siRNA to block the expression of these key genes.
“We go into a cancer cell line with one of our targets of interest and knock out its expression with siRNA then look at the altered phenotype in microarrays to determine what effect inhibition of the target has.”
Avalon avoids so-called “off target” effects, or inaccurate targeting by the siRNA, by making at least four siRNAs and using them for separate transfections and profiles to ensure a “consensus effect,” Dr. Young explains.
When asked if it wouldn’t be simpler to use antisense DNA against known gene targets, Dr. Young notes that, “While in principle antisense and siRNA have the same net effect, we have found that siRNA in general is a more robust technology.”
Verifying Drug Effects
BioImage (Soeberg, Denmark), a biopharmaceutical company focused on cell-based pathway screening spun out of Novo Nordisk, says that using siRNA for pathway profiling studies and to validate their intracellular protein translocation assays has proven very useful.
In BioImage’s assays, the company clones target proteins, fuses them to a reporter systemgreen fluorescent proteinand generates stably transfected cell lines. Automated, high throughput imaging then allows visualization and quantification of test compound effects on the protein’s migrations through the cells.
To verify whether the predicted effect results from the introduction of a test compound, usually a small molecule, BioImage employs siRNAs against effector proteins known to influence the behavior of the target.
Len Pagliaro, Ph.D., BioImage’s vp of business development, explains, “In protein translocation assays, you are essentially monitoring a moving target. In most cases, there are upstream pathway components that affect target protein activity, for instance an upstream kinase. You can introduce an siRNA to the kinase to see if you get the predicted effect of a test compound on the target.”
As an illustration of how this works, BioImage has used its proprietary test system, its Forkhead Redistribution assay, to screen a 250,000 small compound library.
Forkhead transcription factors modulate the expression of a variety of cell cycle regulators, including pro-apoptotic genes. In dividing cells, FKHR accumulates in the nucleus. In growth-inhibited cells, it leaves the nucleus, thereby becoming inactivated, as a result of phosphorylation by the protein kinase Akt.
Compounds promoting the accumulation of FKHR in the nucleus may be expected, then, to inhibit tumor growth. By silencing Akt with siRNA, BioImage scientists could separate its effects on FKHR from those of potential drugs.
BioImage identified a potential anticancer compound series using its Redistribution assay technology and is actively seeking development partnerships.
Finding the Real Target
Kevin Fitzgerald, Ph.D., group leader for high content target validation/emerging technologies in Bristol-Myers Squibb’s (BMS; Princeton, NJ) Pharmaceutical Research Institute, says that the company has been “involved with RNAi from very early on and followed its evolution as a drug development tool.” But, he adds, “We use it knowing full well that mechanistically, RNAi differs considerably from small molecules.”
BMS incorporates RNAi technology into its suite of tools that includes transcriptional profiling, proteomics, and high content screening.
Dr. Fitzgerald also says addition of RNAi technology has helped uncover “several important mechanisms and target proteins for compounds that, while they had unique therapeutic activities in animal models, turned out to have different molecular targets than we originally thought. Early hints from expression profiling showed that even if a small molecule and an siRNA hit the same targets, they have quite different effects.”
For example, he points out, a small molecule may inhibit the activity of one particular protein, such as a tyrosine kinase, affecting its target within an hour. But an siRNA will remove all of the target protein from a cell by eliminating the mRNA from which it is translated, an effect that can take up to 48 hours to occur.
With regard to unanticipated effects, an siRNA to a typical transmembrane tyrosine kinase will not only remove the tyrosine kinase activity of that specific kinase but may result in the disruption of several protein complexes bound to the receptor, or other kinases in that family. A small molecule may result in the loss of only one kinase activity.
“These are different processes with different kinetics,” Dr. Fitzgerald cautions, “So experiments comparing the phenotypes resulting from them are required for drug effect validation.”
Interestingly, Dr. Fitzgerald points out that cell-based assays might come back into fashion. Most companies operate in an ultra-high throughput mode, screening millions of compounds against an isolated, purified target with the hope that when inhibited in the more complex environment of a cell, a potential drug will have therapeutic activity.
But, he says, these methods are unlikely to uncover drugs that are effective because they impact multiple targets, counteracting the fail-safe mechanism inherent in biological systems. He believes that the combination of better cell biology and siRNA may help in this regard.
The reason industry moved away from cell-based screening was due to the difficulties inherent in uncovering molecular targets of the drugs discovered in this fashion. These new technologies now make this easier.
Making Better siRNA Tools
Dharmacon (Boulder, CO) provides tools for RNAi practitioners as well as strategies for minimizing false positive and false negative results. Acquired by Fisher Scientific in 2004, Dharmacon was originally founded to commercialize new technologies for synthesizing RNA oligonucleotides developed at the University of Colorado at Boulder.
Anastasia Khvorova, Ph.D., Dharmacon’s vp of research, talks about the major challenges facing siRNA users and how Dharmacon’s technology has addressed them.
Dr. Khvorova says that most of their customers use siRNA for target validation, knocking out one gene at a time to study pathways and determine the effect on phenotype. Alternatively, she says, many large pharmaceutical clients with greater resources look at broad collections of genes such as groups of kinases, GPCRs, or the current collection of “druggable genes”.
The challenge in performing large screens of this sort is to minimize the number of false negatives that occur when siRNAs fail to downregulate the expression of the intended target gene.
“In principle,” she says, “It’s quite possible to predict how functional an siRNA will be based on its nucleotide sequence. siRNAs function through the RNA Interference Silencing Complex, and some siRNAs are better substrates for this complex than others.”
“While there are exceptions to the rules (some genes are easier to silence than others), certain parameters such as thermodynamic stability profiles are shared amongst all functional siRNA.”
Dharmacon’s approach to the problem of false negatives was to develop a highly predictive algorithm that screens target sequences for traits that are common amongst functional siRNA.
“We use proprietary algorithms to identify the best siRNA target sites within any given gene sequence.” The four top-scoring siRNAs are then provided as a pool (as well as individ ual duplexes) to each customer.
According to Dr. Khvorova, the second major challenge in using siRNAs is minimizing false positives. She says that that the “highly specific” nature of siRNAs is something of a myth. Partial sequence homology between an siRNA and unintended targets can lead to the down-regulation of dozens of genes, and expression changes of these “off-targets” can result in false positives during phenotypic screens.
The problem of off-targeting can be partially solved by mimicking nature. Inside the cell, long double-stranded RNA is processed by Dicer into pools. While the total concentration of siRNA directed against the primary target is high (and thus produces strong silencing), the concentration of any individual duplex within the pool is low, thus minimizing the potential for off-targeting.
Dharmacon has adopted the pooling strategy in their attempts to eliminate off-targeting, and further addressed the problem by identifying a group of chemical modifications that can be applied to any siRNA to increase specificity.
Developed in collaboration with Merck and Rosetta and reported at the recent “Keystone Symposium on RNAi,” Dharmacon believes it has created a “new generation of siRNA reagents”.
Their recently developed Reverse Transfection (RTF) technology will allow rapid, accurate, and cost-effective high throughput knockdown of every gene in the human genome, according to the company.
Dharmacon says that its RNAi product innovations should enable researchers in academic and small biotechnology laboratories to join the ranks of the big pharma labs in being able to cost-effectively perform genome-wide phenotypic screens with minimal or no false positives.
siRNAs as Therapeutics
While companies presenting at “RNAi for Pathway Analysis” will focus on RNAi technology for drug discovery, development of siRNAs as potential therapeutics has moved ahead.
Drug developers cite the need to overcome several obstacles to using these molecules as drugs, including delivery, appropriate targeting, stability, and safety.
Already in Phase I trials with siRNAs targeted at age-related macular degeneration (AMD) are Acuity Pharmaceuticals (Philadelphia) with its Cand5 siRNA drug and Sirna Therapeutics with its siRNA-07. Both drugs target VEGF targets and are designed to shut down the pathological angiogenesis associated with AMD.
Alnylam Pharmaceuticals (Cambridge, MA) also plans to enter the clinic with their siRNA candidate molecule for AMD in the second half of 2005. Reasons for the popularity of this target include a well-validated drug target (VEGF), relative ease of access to the eye, and previous clinical experience and success with ocular antisense drugs, as well as the fact that a successful AMD drug would meet an significant unmet medical need.
According to Zachary Zimmerman, Ph.D., Alnylam’s director of external alliances, the company has addressed several issues impacting the use of siRNAs as drugs. “We know that certain sequence motifs in the siRNAs stimulate the immune system so we avoid these motifs when we make our drug candidates,” he said. “We test a fairly large number of molecules in vitro, and always come up with a handful that will work well without unwanted stimulatory effects.”
Further, Alnylam scientists developed molecular modifications that get around siRNA delivery and stability issues.
In the November 2004 issue of Nature, company scientists described siRNAs stabilized with partial phosphorothioate backbones and with 2-O-methyl sugar modifications on the 3 end to prevent their degradation by endogenous enzymes.
Conjugation with cholesterol facilitated siRNA uptake in target cells of transgenic mice expressing human apoB lipoprotein, resulting in decreased plasma levels of apoB and reduced total cholesterol at a level similar to cholesterol-lowering statins.
The company also recently described results of preclinical in vivo studies showing that a single low dose of the company’s Aln-RSV01 siRNA drug effectively treated respiratory syncytial virus (RSV) infection in mice and protected them against subsequent infection.
Directed against an RSV gene required for viral replication, the siRNA was administered intranasally. Alnylam is developing a formulation of the compound to allow its direct introduction to the lung by aerosolization through a nebulizer.
Alnylam has two alliances with Merck including one for the development and commercialization of its AMD drug, licensing agreements with Isis Pharmaceuticals, and a partnership with Medtronic, as well as collaborations with Mayo Clinic and the Cystic Fibrosis Foundation.
According to Dr. Zimmerman, “We are interested in direct RNAi therapeutics to treat CNS diseases, and we realize that a drug-device combination will be required for delivery. Partnering with a leading medical device company such as Medtronic will help overcome the CNS delivery challenge.”