At A Glance
Name: Ken Reed
Director of Research and Technology
Background: Director of Queensland Agricultural Biotechnology Center — 1991-2000 PhD, Biochemistry, Australian National University — 1974 MS, Biochemistry, Australian National University — 1969 BS, Biochemistry/Chemistry, Melbourne University — 1966
Benitec is a publicly traded Australian biotech firm that has distinguished itself in the RNAi field by earlier this year obtaining the only issued patent — known as the Graham patent — related to the technology aside from the fundamental Fire-Mello patent. With patents in hand, and a licensing deal signed for its technology outside of therapeutics with Promega, Benitec is in the process of becoming a California-based company and is preparing to float its shares on the NASDAQ next year.
Coming along to the US is Ken Reed, the company’s director of research and technology, who recently spoke with RNAi News about Benitec, its technology, and its outlook.
How did you get your start in RNA interference?
In the late nineties, one of our scientists in the Queensland Agricultural Biotechnology Center, a scientist by the name of Dr. Mick Graham, was very inter- ested in the phenomenon of co-suppression in plants and he personally believed that a similar phenomenon operated in mammals — in fact in all complex organisms. The point was, it was one thing to try to establish whether the phenomenon did operate in mammals and another to determine how best to trigger the phenomenon in whatever organism.
There were two elements to what was being undertaken by him at that time: one was, “does this really function in mammals” and “if so, how can we demonstrate it.” Secondly, “how can we trigger it reproducibly and predictably.” The work he undertook had those two outcomes. Firstly, in the late nineties, we observed what you would call homology-induced gene silencing in mammalian cell culture. A lot of work went into trying to work out how best to trigger it, rather than it being a sporadic-type phenomenon as originally described in 1990 in plants — how to make it predictable and usable, in a sense. That work led to initially the filing of patent applications in March 1998, and subsequently we’ve had the first US patent granted for RNA interference in mammals in June of this year.
The work subsequently was on establishing that this was a universal phenomenon, applicable to all species, all cell types, and all genes. We did a lot of work on just extending the initial observations to a number of different systems, and that carried through until about 2001 when the rest of the world finally understood that, yes, this is a characteristic of mammalian cells. But because of the other unique cell-autonomous defenses of mammalian cells, it needs to be treated with care, in a sense, so that these other responses can’t overwhelm the RNA interference response. …
By the late eighties, a company had been set to finance and develop the technology developed in QABC initially. This company was originally known as Ag-Gene. Subsequently, it changed its name to Benitec. In 2000, I left QABC and took a position as research director at Benitec, where I’ve been ever since.
Benitec’s technology is called ddRNAi. Can you give an overview of the technology?
Our starting point was experience with co-suppression in plants, and in plants, by necessity, [co-suppression] must be studied via transgenesis, that is, stable integration of DNA constructs. …
[ddRNAi] is a similar type of approach we started using with animal cells. In this, rather than delivering double-stranded RNA directly to cells or organisms, one designs DNA constructs that will, when transcribed within target cells, yield double-stranded RNA. Now, the constructs must effectively have a promoter, the template, and a terminator. And, I guess, one of the great advantages of the approach we finally adopted was to have the transcript run off as a single molecule, a hairpin-type structure, which is self-complimentary. … [This] means the construct is an inverted repeat with a small space in between the sense and the antisense strand.
This has a lot of advantages. For a start, when a transcript is run off as a single molecule, you get very efficient, very fast, and complete annealing of the complimentary strands, what we call snap-back annealing. But it turns out that in mammals, in particular, this is a very good way to go, in that mammalian cells, in particular, don’t like double-stranded RNA and they have a number of defense mechanisms to protect against it. ... But it turns out that if … double-stranded RNA is transcribed within a cell, rather than presented to it from the outside, those responses are far more muted. …
[Additionally], one can use a variety of delivery vehicles to get the material into the cell. In terms of the therapies, the most common will almost certainly be viral vectors, where there are half a dozen types available to achieve virtually any specificity and penetration you wish. It adds enormously to the scope of how one can deliver RNAi to mammalian cells. … In the words of Phil Sharp, the whole issue with siRNAs is delivery, delivery, delivery. We think that using viral vectors as an option for delivery does indeed bring RNAi therapy into the realm of the very, very possible.
Another advantage is that one can control the extent of RNAi within particular cell types through the use of suitable promoters. One can use tissue-specific promoters to further enhance the target specificity.
There are other advantages that are more on the path to therapy development rather than the use of it as a therapy, per se. One of the greatest is in the production of transgenic animals. One can make an RNAi construct a part of the genetic material of a model animal, for example, and that allows you to then do your target validation in a complete systems biology approach. … If you prepare a transgenic animal with an optimal construct, you can examine the effects of that construct, that particular RNAi-inducing material, in the whole organism … you’re getting absolute disease modeling and target validation, but more than that, in validating the target, you’re also validating a potential therapeutic agent. … So we’re now, in a few years, talking about effectively eliminating that entire small molecule screening aspect of drug development, which has been the bane, on one hand, of the drug industry.
Some people have expressed concerns over the use of viral vectors. Are you aware of these concerns?
Oh, gosh, yeah. They derive from the small number of adverse events seen in gene therapy trials … They are not general concerns in gene therapy, rather … specific to the cases. There’s no question, viral vectors for gene therapy are in much better shape that they ever have been. But more than that, there’s a bit of range available. Obviously, its not something anyone’s going to go into gung-ho, but there are so many trials around and so many improvements … that we’re very optimistic about the outcomes. …
Gene therapy as we presently think of it is used to deliver a protein to appropriate target cells in amounts sufficient to relieve a symptomatic deficiency in a patient. In most cases it has proven extremely difficult to achieve this. [This is] hardly surprising, since this is a very difficult undertaking. Delivery to appropriate cells of a gene encoding the protein is just the first step. That gene must be transcribed efficiently, the RNA transcript must be processed into a functional mRNA and exported from the nucleus to the cytoplasm where it must be translated into a functional protein, which may itself need to be modified and localized to achieve its effect. All steps must work efficiently and the effect must be sufficiently long-lived to provide relief from the disease symptoms. This type of gene therapy is a wonderful feat of bioengineering and the fact that there have been successes is quite remarkable.
Gene therapies that will be used to deliver ddRNAi do not have such stringent constraints. The package that must be delivered is smaller, allowing flexibility in choice of delivery agent. [Also,] the range of promoters available for driving transcription is wider, including pol III in addition to pol II; efficient transcription is essential, of course, but with a smaller, defined transcript. Nuclear processing and export are similarly definable, with alternative pathways available dependent on the promoter used.
The major technical difference is that the exported small dsRNA is itself the effector — it is not translated into protein, needing only to engage RISC to trigger RNA interference, either directly or after it has been cut by Dicer.
A further difference is in the purpose of the therapy. Traditional gene therapy seeks to replace a missing function. ddRNAi therapy seeks to eliminate a harmful function, presenting a large number of defined and definable targets. Obvious therapeutic targets for ddRNAi include cancers and chronic viral diseases such as HIV, AIDS, and hep C. With cancers the intent of the therapy is to kill cells so the treatment need not necessarily be long-lived. With chronic viral diseases this would seem to be advantageous but we will not know the importance of longevity until trials are undertaken. In cancers and viral diseases in particular, ddRNAi has the advantage of allowing multiple gene targets to be hit simultaneously, reducing (or possibly eliminating) the probability of selection for resistance. This capacity for multiple hits also suggests the future possibility of ameliorating multi-factorial diseases. Also … ddRNAi lends itself to adjuvant therapies, for example in combination with antibody or chemotherapy and perhaps in [combination] with traditional gene therapy to replace a harmful protein with its normal counterpart.