In the development of RNAi-based therapeutics, delivery is undoubtedly the number one challenge, at least right now. But as researchers work on clearing this hurdle, the question of how best to determine in vivo activity of these compounds looms in the shadow of antisense drug failures.
“It’s important to set a standard in the industry for what is a demonstration of in vivo activity for an siRNA because it’s so much the question. When people make claims about in vivo activity, we think it ought to be corroborated with the right level of standardized data,” Alnylam CEO John Maraganore told RNAi News.
To do so, Maraganore said, would be to learn an important lesson from antisense development. “A lot of the biological effects that we were seeing with antisense oligonucleotides were non-specific effects based on activation of innate immunity.”
Intradigm’s CSO Martin Woodle agreed. “In the antisense field, a lot of effects were observed, but they turned out not to be necessarily antisense effects,” he told RNAi News.
To sidestep this issue, Woodle said that Intradigm looks for three things to gauge siRNA activity: a reduction in mRNA levels, a reduction in protein levels, and, of course, the desired phenotype effect.
“We tend to use phenotypic effects first, then … when we obtain that, we go on to characterize either [mRNA] or protein level, or both,” Woodle said. “The other parameter one can use is looking at sequence selectivity, how many sequences give the effect on any one of those three signals. If it’s reasonably good gene selectivity, one ought to get activity only out of active sequences.”
But, even these criteria may not be enough, Woodle said. “From our standpoint, three signals alone doesn’t provide the strongest evidence. Seeing message[nger RNA] or protein levels with phenotype needs to be also shown with control sequences not giving those effects.”
The use of control siRNAs is key when doing experiments where evaluating mRNA or proteins levels might not be feasible, said Dmitry Samarsky, director of technology development for Sequitur, which has an ongoing RNAi collaboration with Intradigm.
“Unfortunately, it is not always easy” to demonstrate that a transcript and protein have been knocked down, Samarsky told RNAi News. He gave the example of delivering an RNAi compound to cells that makeup only about 30 percent of a particular organ. “If you try to monitor the transcript knockdown, you isolate the whole organ … and you cannot see more than 30 percent knockdown, anyway, because 70 percent of the cells will not have the knockdown” since the transcript is not expressed in those cells.
“That’s why, in many cases, people seem to go further down to the phenotype” alone, he said. “If you … only have access to the phenotypic stage, then you have to be very rigorous in providing controls, like several non-targeting [siRNAs] versus several targeting [siRNAs]. That way, every time you have a nice correlation, that would be a strong indication that what you see is what is happening.”
In the end, though, an effect on mRNA and protein levels, as well as a phenotypic effect, seems to be the gold standard.
“[In vivo activity] requires a demonstration of a statistically significant reduction level of messenger RNA levels, it requires statistically significant reduction of the corresponding protein levels, and it requires correlative data on pharmacokinetics and biodistribution,” said Maraganore.
Samarsky added that “something that we sometimes find a little bit funky is when people say ‘Oh, we just go and do Western [blot analysis]. Isn’t [phenotypic activity] the ultimate goal?’ Yeah, it’s the ultimate goal, but [reliance on just the phenotype] isn’t rigorous science.”
Despite successes with these in vivo markers, however, more work needs to be done.
“This is like the shore of an ocean,” Samarsky said. “All we know is the little bit we see, but the ocean could be huge.”
—DM