Vice President of Molecular Genetics
Integrated DNA Technologies
At A Glance
Name: Mark Behlke
Position: Vice president of molecular genetics, Integrated DNA Technologies
Background: MD, Washington University, St. Louis 1988; PhD, immunogenetics, Washington University, St. Louis 1988; BS, biology, Massachusetts Institute of Technology 1981
Though trained as a physician, Mark Behlke had always intended on pursuing academic research, following up his residency at Brigham and Women's Hospital with a postdoctoral stint at the Whitehead Institute. When the opportunity to enter industry arose, however, he took it.
Now at IDT, he has had to opportunity to collaborate with colleagues at the City of Hope's Beckman Research Institute on the development of 27-mer RNA duplexes that appear to trigger an RNAi effect more potent that standard 21-mer long siRNAs. Recently, Behlke spoke with RNAi News about the 27-mers.
Could you give a little background on the 27-mer technology and how it was developed?
The whole project started out as a collaboration between John Rossi's lab at the Beckman Research Institute at the City of Hope in Duarte, California, and my lab here at IDT. One of John's postdocs, Dong-Ho Kim, had an observation that some [siRNA] samples that had 5' overhangs were more potent in triggering RNAi than some of the compounds that they were originally using. So we embarked on a study to try to understand this effect better. It turns out it wasn't truly 5' overhangs that were having the effect rather, it was that the constructs they had made that had 5' overhangs were longer.
As we explored this further, we made a series of oligos that walked out in length and found that at select sites in our target gene we were studying EGFP there were cases where the longer oligos were as much as 100-fold more potent on a molar basis, which surprised us quite a bit because nothing like this had been reported before. I think that a lot of work in this area was somewhat biased by the incredibly good quality data that had come out Tom Tuschl's group earlier where, at least in Drosophila, things that were 25 bases and longer in the short range didn't work at all. But the biochemistry is different between flies and humans, and it had been shown as early as 2001 that longer oligos did work in mammalian systems, but I think that was largely ignored. And in fact, in our own work, if we hadn't been doing careful dose response curves, we never would have seen a difference either because if the 21-mer was working, and the 27-mer also worked, we never would have noticed the difference except we were doing careful dose response curves.
At the time, we were starting with a relatively high dose, 50 nanomolar, as was common practice at that time two years ago and we were looking down in the sub-nanomolar range, down to as low as 200 picomolar or 50 picomolar. That's where we saw the differences. As you dropped the dose, the longer sequences retained their potency better.
All this was done in vitro?
Oh yes. All the earlier work was done in cell culture transfection studies.
The first sites that we looked at we saw as much as 100-fold increase in potency, and it turns out that that was a two-fold effect. It was additive contributions of the 27-mer being processed into a better 21-mer in addition to some effect that was 27-mer-specific. So each one of those individual contributions gave about a 10-fold boost, which added up to the 100-fold boost.
Do you have any sense of what some of these 27-mer-specific effects are?
Our best guess at that is there is some linkage of the Dicer processing into RISC loading. Now, in Drosophila, there is an absolute linkage, and if you knock out Dicer-2 or R2D2, which are the co-factors involved in processing cleavage of double-stranded RNAs, you can't get even 21-mers to work. So in flies you have to have Dicer to load RISC.
In humans it's much more complex; there's only one Dicer, while Dicer knockouts [are] not viable as animals, you can use 21-mers in Dicer knockout cells. The equivalent of the R2D2 ortholog for humans is TRBP, and that has been identified. This field of study, at this juncture, is relatively new, and so the whole story isn't out on that although there is some suggesting that if you knock our TRBP the RISC loading is less efficient. But it's clearly not needed like it is in flies, so if you knock out Dicer and knock TRBP, you can still have RNAi working in mammalian cells.
So what we think is perhaps happening is you have the option in mammalian cells to have some sort of facilitated protein complex loading akin to what happens in Drosophila, or that you can simply have the 21-mer dissociate from Dicer after cleavage is done and have that load into RISC without having to have participation of the other protein components. [We also think that] possibly some element of this facilitated loading into RISC is what's coming into play in terms of making some of these longer sequences a little bit more potent.
Do you find these work more effectively in certain cells and you're having harder times in others?
There's definitely a cell type-specific effect as far as how potent RNAi is. I don't know if any differences we have in potency for 27-mers is any different than potency for 21-mers. Certain cell lines definitely appear to be more potent for RNAi effect than other cell lines.
Have you explored modifications [on the 27-mers]? I recall you saying [during your presentation at this year's RNAi Europe conference in Amsterdam] that what works for 21-mer siRNAs isn't necessarily what works for 27-mers.
We're in the process of doing those experiments right now. The one thing we do know is that if we block the ability of these to be diced … they don't work. So our compounds need to be processed by Dicer, which obviously changes the modification pattern that you can use. A lot of … modifications … are being … avidly promoted today by companies like Sirna or Atugen, where there've been some really nice publications done about the benefits of chemically modified duplexes. These are heavily modified duplexes, and some of these have almost no RNA in them they're all modified. Obviously that's not going to work for us because we have to remain dice-able. So I think that for in vivo use, the kind of modifications we're going to employ are going to be a little bit different from some of the routes taken by these other groups because we're going to want to focus more on the minimal modifications that will improve compound stability to whatever extent is necessary, as well as modifications that may alter the ability of the duplexes to trigger the innate immunity system.
What about in vivo use? Is there any work going on trying these things out [in animals]?
Very much so. We have one group that is using the simple hydrodynamic injection that was developed by Mirus just to demonstrate that these work in mouse livers. This is being done with injection of a luciferase plasmid with the RNAs, and the 27-mers are working quite well.
That's being done at IDT?
No, that's being done at CalTech. Mark Davis and Jeremy Heidel, who were co-authors on the recent Nucleic Acids Research paper [on the use of 27-mers], are doing that work.
[Additionally], Ed Cantin at the City of Hope has got a mouse TNF system, where they're studying the effects of TNF both in terms of herpes virus infection, as well as the main cytopathic mediator of the shock effect that comes out of LPS injection. Ed is studying the use of the 27-mers in blocking the TNF response to LPS or HSV. In that system, we're using intraperitonial injection with the Mirus TransIt-TKO lipids.
What do you think about the therapeutic potential of these 27-mers?
I think that the therapeutic potential will be similar to the 21-mers. It's just going to be a different pattern of chemical modification. When it comes down to getting a good therapeutic, if the 27-mer conversion can even just double potency, that might have some real benefit. We're hoping that when it comes down to therapeutics, these compounds will have some advantage.
Again, though, it's going to require that the delivery system provides some of the protection for the molecule. I think that's going to be a common theme, though the groups that want to use fully modified RNAs, or things that aren't even RNAs any longer, compared to the groups that want to use the delivery system to protect the payload and have a less modified oligo. Having a less modified oligo is going to have advantages and disadvantages. The advantages will be that they'll be less expensive to make and they'll be more physiologic; if you can limit yourself to modifications that are naturally occurring, there may be fewer toxicity issues.
What about on the research side? What are the key benefits for the research market to use these things?
One of the things we're finding [is that] people are doing is they're making 27-mer conversion for targets they really care about because you can use less of these. I think dosing is getting to be an issue people are paying more attention to now that they're becoming more aware of off-target effects and the like. So the ability to use smaller amounts is advantageous.
We [at IDT] have not taken the track of making these pre-made, pre-packaged whole-genome libraries, or large subset libraries, that some other companies have made, mostly because I don't think there is any benefit to having yet another library on the market. I think there are quite enough out there already. But people, once they've used [a whole-genome] library and identified an interesting target, … are coming in and getting 27-mers for that target and finding some benefit to having a higher potency reagent.