Looking to cash in on the market for RNAi libraries, a small biotech company called SomaGenics has begun developing siRNA libraries that are specifically designed to increase the ease with which scientists can select the most effective gene expression inhibitors.
“We focused on the need to make combinatorial libraries of inhibitors based on nucleic acids and select from those libraries therapeutic molecules that work best within the cell,” Brian Johnston, president, CEO, and founder of SomaGenics, told RNAi News.
“To know which members of your library work best, you should have nature select out the ones that do what you want inside the cell,” he said. “If you can select for the molecules that are superior in the in vivo setting, in cell culture, then they have a higher probability of working in humans [rather] than molecules that are tested in a non-cell environment.”
According to Johnston, a former researcher at SRI International, problems can occur when researchers begin searching for the siRNAs (or other gene-silencing molecules, for that matter) that work best.
“If you start out with a completely randomized library, then you have a problem of numbers,” Johnston said. “For antisense sequences that are 20 nucleotides long, or thereabouts … then the number of different molecules you have from a random sequence is four to the 20th power — about a trillion. If you want to test those in cells, where you have one species per cell, you need a trillion cells, which is not really practical for most of us.”
To simplify things, SomaGenics is developing what Johnston terms “directed libraries” of siRNAs.
“Instead of all possible sequences of a given length, we have all possible sequences contained within a target of a given length,” he said. “For example, if we had a 10,000 … nucleotide [long] messenger RNA, one member of the library would be nucleotides one to 20, another would be two to 21, three to 22, and so on — it would be like walking down the message.
“All possible 20-mers in that sequence would number just under 10,000, which is a whole lot less than a trillion,” he added. “You could then insert this library into cells, begin to screen the cells and recover those that have the phenotype that you’re after, and analyze for the sequence.”
Johnston said that directed libraries could be made to target a specific mRNA, a family of mRNAs, or even an entire genome. “In these methods for making the libraries [that] we’ve developed, it all depends on the starting material you feed into the process,” he said. “If you give it as a target a single RNA, it will give you all sequences within that single molecule; if you give it a whole collection of molecules, it will give you all sequences within all of the members of that collection.”
Johnston said that this approach keeps the number of siRNAs dealt with to a reasonably low number. “Even within the human genome, the complexity of your library is going to be at most a billion or so — and realistically tens to hundreds of millions, which is still a manageable number of cells if you’re willing to work with a number of plates.”
Johnston said that SomaGenics has just started testing out the libraries in cells. The company has also just received a one-year NIH grant worth between $250,000 and $300,000 to develop directed siRNA libraries targeting dengue virus.
The grant project, which is set to run between April 1 of this year and the end of March 2005, involves linking potential target dengue viral sequences to the HSV thymidine kinase gene, and then expressing them in stable human cells, according to the grant’s abstract.
A directed siRNA library, designed to target only dengue sequences, will be introduced into the cells. Ganciclovir, which kills any cell expressing the HSV-TK protein, will then be administered to the cells, the abstract states. Cells that survive the ganciclovir will be isolated and the rescued siRNAs will be identified.
Aside from its library work, SomaGenics is also developing a novel antisense-related gene-silencing technology called RNA lassos.
According to Johnston, RNA lassos are single-stranded RNA molecules that contain a “sequence complementary to a target, typically a messenger RNA. [The molecule] is linked along its chain to a ribozyme whose function is to link the ends of the molecule together so that it can interconvert from being a linear molecule to a circular molecule,” he said.
Once the RNA lasso binds to its target, Johnston said, the linked ends form a knot with the helix. “The benefits of the knot are that it’s much more difficult to displace by, say, the translation machinery … and [is] resistant to conditions that would normally displace it, such as higher temperature.”
SomaGenics is exploring the development of RNA lassos as therapeutic molecules, and preliminary work has been conducted using the molecules to target tumor necrosis factor-alpha. Johnston said that the company is considering using the molecules for autoimmune and inflammatory conditions, including psoriasis.
Johnston added that he hopes SomaGenics would be able to start a phase I trial of an RNA lasso-based drug in a dermatological indication within a couple of years.