While an siRNA’s ability to target a specific gene is key to its development as a therapeutic, it is also important to consider early in the drug-development process whether the siRNA has off-target effects resulting from seed region sequence complementarity, according to a senior Rosetta Inpharmatics official.
Peter Linsley, executive director of research at Rosetta, said such a move could help reduce unexpected toxicities that might not surface until the clinical testing phase.
“We can get clean siRNAs if we work at it, [ones] that are less susceptible to these sequence-dependent effects,” Linsley, who spoke last week at the third annual Oligonucleotide Therapeutics Society meeting in Berlin, told RNAi News. “What we’re trying to do is look earlier in the process to get the cleanest siRNAs we can at the beginning of the process rather than waiting until the end.”
At the meeting, Linsley offered a glimpse at some experiments, conducted on behalf of colleagues at Sirna Therapeutics, showing how siRNA that seemed highly specific in mice could hit unintended targets once evaluated in human cells.
Both Rosetta and Sirna are subsidiaries of Merck.
To arrive at those findings, Sirna provided Rosetta with four chemically modified siRNAs targeting an undisclosed gene and asked, “which would be the best to pursue in development,” Linsley said during his presentation.
All four siRNAs were highly effective in knocking down their target gene and appeared to cause little unintended silencing when examined in mice. But when the siRNAs were evaluated in human cells, “two had really bad liabilities,” he noted.
“We found that the transcripts regulated by [one] particular therapeutic candidate overlap quite strongly with miR-30a-3p,” he said. “I cannot tell you that this will cause a problem … but why would we want to [take a chance and] put this into people?”
Microarray analyses showed that the second siRNA, meanwhile, hit both its target gene and a gene associated with a liposomal storage disorder.
“Again, I can’t tell you this will cause a problem, but it doesn’t seem like a good idea” to use this in humans, Linsley said. “Our data suggests that cultured human cells are better models than live mice for [evaluating] seed sequence-dependent off-target silencing.”
He said if researchers are regulating a set of transcripts in human cells that have an undesirable effect, “those regulations won’t be present in mice … because the target sites don’t exist in mouse transcripts.
“You’ll see other things — maybe knockdown of your target is not good for some reason … and you may or may not see any immune effects — but you won’t see sequence-specific effects in mice because they won’t be the same in mice as they would be in humans.”
To Linsley, one way of overcoming this issue is to conduct experiments in human cells to identify possible specificity problems earlier in the siRNA-based drug-development process.
Linsley stressed that examining specificity early on won’t make toxicology studies any less important, “but why wait [to do toxicology work on] something that has a big spectrum of off-target effects” when you can identify an siRNA with few off-target effects at the beginning?
He also told RNAi News that he doesn’t see much value in toxicology work in mice when it comes to looking at off-target effects tied to seed region sequence complementarity, and that conducting these kinds of experiments in primates after examining an siRNA-based drug candidate might be a better option.
Such measures could help the RNAi drug field avoid the kind of problem that arose when TeGenero Immuno Therapeutics tested an anti-CD28 monoclonal antibody in a phase I trial last year, Linsley said.
“What we’re trying to do is look earlier in the process to get the cleanest siRNAs we can at the beginning of the process rather than waiting until the end.”
According to a report published in The New England Journal of Medicine, six healthy volunteers enrolled in that study experienced multi-organ failure after receiving a single dose of the antibody due to a sudden and rapid release of pro-inflammatory cytokines.
TeGenero “brought these antibodies forward and they nearly killed some people because they didn’t have the appropriate [toxicology] species,” Linsley said. “I think the same situation is true here. What it says is that the real [toxicology model] is going to be the species as close as possible to the species you’re intending to go into, [which is] people.”
Whether companies developing siRNA-based drugs will conduct primate toxicology studies so early in the drug-development process, or even be required to by regulators, is uncertain.
As Linsley noted, “it’s a lot easier to do [this work] in mice.”
In 2003, researchers from Rosetta published in Nature Biotechnology data showing the potential for certain siRNAs to silence non-target genes. Specifically, 16 siRNAs that had been designed to target the coding region of the IGF1R gene in HeLa cells, and eight that targeted the MAPK14 gene, actually regulated different genes.
During his presentation at the Oligonucleotide Therapeutics Society, Linsley highlighted this work, which caused a stir within the nascent RNAi field, and presented the newest off-target data as a sort of follow-up to the earlier research.
But despite being well-received by the conference attendees, the new data may not be published, he said.
“I don’t think there is anything secret about [this work, and Merck] would probably let us publish it,” Linsley said. “I just can’t imagine what journal would take this.”