Researchers from the University of Amsterdam this month reported new details in Nucleic Acids Research on how a particular shRNA configuration is processed through a Dicer-independent pathway while still maintaining RNAi properties, opening the door to a new shRNA design with enhanced therapeutic properties.
Specifically, the team found that shRNAs that are shorter than traditional short hairpins and maintain a relatively small terminal loop size are processed by Argonaute2. Importantly, the resultant molecules — dubbed AgoshRNAs — have only one active strand, with the passenger strand destroyed by Ago2.
According to Ben Berkhout, the study's lead author, this feature could potentially lead to the creation of therapeutic shRNAs with fewer off-target effects.
The work, published in Nucleic Acids Research, stems from Berkhout's longstanding interest in developing an shRNA-based treatment for HIV/AIDS, which has thus far yielded a number of drug candidates that have performed well in vitro and are being examined in vivo.
He said that as part of that work, his lab has explored different modifications to HIV-targeted shRNAs in an effort to improve their properties.
“In the past, we increased loop size and increased the nucleotides of the loop, and at some point we were making the stem region of the short hairpins a little bit longer or a little bit shorter,” he told Gene Silencing News. “As expected, if they become too short, then the [RNAi] activity is gone.”
At some point during this experimentation, Berkhout and his colleagues began looking not only at the activity of the shRNA's guide strand, but also the passenger strand. They discovered that, in some cases, the former lost gene-silencing activity but the latter gained it.
“We did Northern blots to see whether the RNA was perhaps processed differently and saw a completely different picture” than what was anticipated, he said.
When an shRNA is normally processed by Dicer, the result is two strands of roughly 21 nucleotides. “But in this case, when the hairpins were too short, we saw a much longer product ... of about 35 nucleotides,” which is not consistent with standard Dicer processing.
Additional experimentation confirmed the effect with AgoshRNAs with different step sequences.
With further examination, Berkhout's team saw that cleavage was occurring halfway down the 3' side of the shRNA duplex, which “reminds you of what Ago is going with RNA duplexes” when it cleaves target mRNA or the passenger strand of an siRNA duplex between nucleotides 10 and 11 from the 5′ end of the guide strand, he noted.
“We think that if we make the hairpins too small, they cannot be recognized by Dicer,” Berkhout explained. “Then an alternative processing pathway opens up,” taking advantage of Ago2.
This phenomenon had previously been identified by others, notably by SomaGenics CEO Brian Johnston, but the exact nature of the processing mechanism was not identified.
“They were very close … but could not identify Ago,” Berkhout said.
Interestingly, he said that the Dicer-independent processing of AgoshRNAs is similar to what has been observed with certain microRNAs, particularly miR-451, as described in a number of recent publications.
“Mechanistically, it looks like the same thing is happening” with AgoshRNAs as with this subset of miRNAs, he noted.
Overall, to take advantage of Ago2 processing, shRNAs must have a stem length of around 19 base pairs, Berkhout said. Additionally, their terminal loops “should not be too big [because] then you create problems of the short hairpin getting into Ago in the first place.”
While scientifically interesting, he also sees real-world benefits of AgoshRNAs.
“The way this molecule is processed, you're left with only one active strand,” reducing the chance of off-target effects, Berkhout said.
Further, AgoshRNAs form smaller hairpin duplexes than regular shRNAs, and may “exhibit a better safety profile concerning activation of the dsRNA-induced protein kinase R/interferon pathways,” he and his colleagues wrote in Nucleic Acids Research. At the same time, “because AgoshRNAs do not mature via Dicer, they also do not compete with this aspect of miRNA biogenesis.”
The scientists also noted in their paper that Ago2-mediated processing may yield “more distinct RNA molecules as Dicer creates imprecise ends,” and that they could prove to be attractive gene-silencing molecules for use in Dicer-deficient cells.
Experiments are ongoing testing the safety of AgoshRNAs and “whether conversion of the regular loop sequences into anti-HIV sequences will generate more potent and possibly escape-proof antivirals,” the researchers concluded.